JP2011141777A - Apparatus, method and program for data summarization - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To continue data summarization, even if a target period for data summarization is extended when a data value is suddenly changed. <P>SOLUTION: The data summarization apparatus for summarizing successively generated data includes: an all approximation formula memory for storing a plurality of approximation formulas for approximating data; a representative approximation formula selector for selecting an approximation formula in which an approximate value found out by using each approximation formula based on present time is included within a range scope to be a scope of values to be taken by data from the approximation formulas stored in the all approximation formula memory, as a representative approximation formula to be a candidate for approximating newly generated data; and a representative approximation formula memory for storing the representative approximation formula selected by the representative approximation formula selector. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a data summarization apparatus, a data summarization method, and a data summarization program that reduce the amount of information by sequentially summarizing sequentially generated data by approximation.

  There has been proposed a technique for approximating the value of time series data by representing it by a plurality of approximation formulas. For example, Patent Document 1 describes a related technique for creating a function that approximates learning data based on the learning data.

  In the related technique described in Patent Document 1, learning data is first clustered, then an approximate expression representing the data of the cluster is created for each cluster, and an error in the approximate expression of each cluster is evaluated. Then, clustering, approximation formula creation, and evaluation are performed recursively for clusters with large errors in approximation formulas. Further, the cluster information and each approximate expression are stored, and the input data first determines the cluster to which it belongs, and the approximate value of the input data is calculated using the approximate expression of the cluster.

  For example, Patent Document 2 describes a data compression and storage device that approximates and stores plant data using an approximate expression as a related technique. Further, in the related technique described in Patent Document 2, the deviation between the value predicted by the approximate expression and the actual input data value is obtained, and if the obtained deviation is larger than a predetermined threshold, the input data is stored. It is described that it is determined that the data is stored, and if it is smaller than a predetermined threshold, it is determined that the data need not be stored.

JP-A-7-65168 Japanese Patent No. 3242936

  In the related technique described in Patent Document 1, clustering and creation of an approximate expression are recursively repeated until the accuracy reaches a certain level. Therefore, in the case of approximating data that occurs sequentially and whose value changes suddenly, the number of clusters and approximate expressions continues to increase. Then, by applying the related technology described in Patent Document 1 to a data summarization device that approximates continuously generated time series data with an approximate expression, and continuously handling data that occurs continuously over a long period of time, The amount of calculation of cluster determination performed each time data is generated continues to increase. If the amount of calculation for each data generation continues to increase and exceeds the processing speed of the data summarization apparatus, the approximation of the data by the approximate expression cannot be performed.

  Therefore, even if the related technique described in Patent Document 1 is used, if the data value that occurs sequentially changes suddenly, the data summarization is performed when the period of data summarization becomes longer. It becomes difficult to continue.

  An example of data that is sequentially generated and whose data value changes suddenly is a CPU usage rate. As described above, the related art described in Patent Document 1 is not suitable for compressing data whose data values suddenly and suddenly change greatly.

  In addition, by using the related technique described in Patent Document 2, it is possible to determine how far the actual input data value is from the value expected by the approximate expression and to select the data to be stored. , Saving data efficiently. However, approximation of input data is performed using only one approximation formula, and this is only suitable when the change tendency of the value of the input data is constant to some extent. Therefore, it is impossible to approximate data that changes suddenly by using the related technique described in Patent Document 2.

  Therefore, the present invention provides a data summarization apparatus, a data summarization method, and data capable of continuing data summarization even when the data summarization period is long when the data value suddenly changes. The purpose is to provide a summary program.

  A data summarization device according to the present invention is a data summarization device for summarizing sequentially generated data, and is used for approximating newly generated data and all approximation formula storage means for storing a plurality of approximation formulas for approximating data. As a representative approximate expression that is a candidate for, the approximate value obtained using each approximate expression based on the current time from the approximate expressions stored by all approximate expression storage means is a range of values that the data can take And a representative approximate expression selecting means for selecting an approximate expression included in the range and a representative approximate expression storing means for storing the representative approximate expression selected by the representative approximate expression selecting means.

  Another aspect of the data summarization device according to the present invention is a data summarization device for summarizing sequentially generated data, all approximate expression storage means for storing a plurality of approximate expressions for approximating data, and newly generated data As an approximate expression that is a candidate for approximating, an approximate value obtained by using each approximate expression based on the current time from the approximate expressions stored by all approximate expression storage means is a value that can be taken by the data. A representative approximate expression selecting means for selecting an approximate expression within a predetermined range and within a predetermined number, and a representative approximate expression for storing the representative approximate expression selected by the representative approximate expression selecting means And a storage means.

  Still another aspect of the data summarizing device according to the present invention is a data summarizing device for summarizing sequentially generated data, a total approximation formula storing means for storing a plurality of approximation formulas for approximating data, and a newly generated data summarization device. Representative approximate expression selection means for selecting an approximate expression within a predetermined number from among a plurality of approximate expressions stored in all approximate expression storage means as representative approximate expressions that are candidates for approximating data, and representative And a representative approximate expression storing means for storing the representative approximate expression selected by the approximate expression selecting means.

  A data summarization method according to the present invention is a data summarization method for summarizing sequentially generated data, in which a plurality of approximation formulas for approximating data are stored in a total approximation formula storage means, and newly generated data is approximated. As a representative approximate expression that is a candidate for, the approximate value obtained using each approximate expression based on the current time from the approximate expressions stored by all approximate expression storage means is a range of values that the data can take And an approximate expression within a predetermined number is selected, and the selected representative approximate expression is stored in the representative approximate expression storage means.

  A data summarization program according to the present invention is a data summarization program for summarizing sequentially generated data, and is newly generated in a computer having all approximate expression storage means for storing a plurality of approximate expressions for approximating data. As a representative approximate expression that is a candidate for approximating the obtained data, the approximate value obtained using each approximate expression based on the current time can be taken from the approximate expression stored in the all approximate expression storage means. A process of selecting an approximate expression that is included in a value range that is a range of values and that is within a predetermined number, and a process of storing the selected representative approximate expression in the representative approximate expression storage unit are executed. Is for.

  According to the present invention, when the value of data suddenly changes, data summarization can be continued even if the period for which data summarization is performed becomes longer.

It is a block diagram which shows the structural example of the data summarization apparatus by this invention. It is explanatory drawing which shows an example of the data showing the approximate expression which all approximate expression memory | storage means memorize | stores. It is a block diagram which shows the structural example of a representative approximate expression management part. It is explanatory drawing which shows an example of the data showing the range which a range storage means memorize | stores. It is explanatory drawing which shows an example of the data showing the representative approximate expression which a representative approximate expression memory | storage means memorize | stores. It is a block diagram which shows an example of a structure of an approximation means. It is explanatory drawing which shows the example which selects the approximate expression which satisfy | fills the precision restrictions defined beforehand. It is a flowchart which shows the process example of the selection process of the representative approximation formula which a data summarization apparatus performs. 12 is a flowchart illustrating an example of an approximate expression selection process executed by a representative approximate expression first limiting unit. FIG. 6 is a block diagram showing a configuration example of a data summarizing device according to a second embodiment. 10 is a flowchart showing a processing example of representative approximate expression selection processing executed by the data summarizing apparatus according to the second embodiment. FIG. 10 is a block diagram illustrating a configuration example of a data summarizing device according to a third embodiment. It is a block diagram which shows the minimum structural example of a data summarization apparatus.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. First, an outline of the entire data summarizing apparatus according to the present invention will be described. The data summarization apparatus according to the present invention reduces the amount of data that is sequentially generated over time. By reducing the amount of information, it is possible to reduce the storage capacity required for storing data, compared to the case where the data itself is stored exactly as it is. In the present embodiment, reducing the amount of data information in this way is referred to as “summary”.

  In general, although the accuracy of data is reduced by summarization, the data summarization device according to the present invention is data that is generated with a constant tendency sequentially over time, and the tendency of the data may change irregularly. Summarize certain data accurately and efficiently.

  For example, “CPU usage rate” is an example of data that is generated sequentially with a certain tendency and the tendency of the data may change irregularly. In addition, not only the CPU usage rate, but also the information such as the number of accesses per unit time of the web page to be observed, the total number of accesses after the web page has been released, etc. suddenly change, so This corresponds to “data that occurs sequentially with a certain tendency and may change irregularly”. In addition, for example, the communication amount of the network device also corresponds to “data that is sequentially generated in a certain tendency and the data tendency may change irregularly”. “Sequentially generated data” is a concept including “sequentially observable data”.

  In the following description, numerical values that occur over time will be described as an example as “data that is generated sequentially with a certain tendency and may change irregularly”. However, even if the data itself is not a numerical value, it can be converted into a numerical value, and any data can be used as long as the difference between the converted data can be derived.

  In the example shown below, each data includes a data value and an occurrence time of the data value in association with each other. In the following description, the occurrence time of this data value may be simply referred to as the occurrence time of data.

  In the following description, the case where data that occurs sequentially in time series is used is shown. However, if the data has some order, for example, the positional order such as the order of distance or the order of distance It can also be applied to the case of performing analysis by sequentially inputting characteristic data. Further, the data is not limited to data generated continuously at short time intervals (for example, every few seconds), but may be data generated at long time intervals such as several hours or days as long as the data is generated sequentially. It can also be applied to the case of handling data whose occurrence interval is not constant.

  In addition, the data summarization apparatus according to the present invention derives a function for calculating a data value (numerical value) based on a set of generated data, with the generation time as a variable. This function is an approximate expression for obtaining an approximate value of a data value based on the occurrence time.

(First embodiment)
Next, a first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration example of a data summarization apparatus according to the present invention. As shown in FIG. 1, in this embodiment, the data summarizing apparatus includes a representative approximate expression management unit 1, a total approximate expression storage means 2, a representative approximate expression storage means 3, a data input means 91, and an approximation means 92. Including. The data summarization apparatus is specifically realized by an information processing apparatus such as a personal computer that operates according to a program.

  Specifically, the total approximate storage unit 2 is realized by a storage device such as a magnetic disk device or an optical disk device. All approximate expression storage means 2 stores all approximate expressions for obtaining past data values (approximate values) in association with IDs for identifying the respective approximate expressions. For example, when past data values are represented by one-dimensional real numbers and the data values are approximated by a linear function with respect to the time of occurrence of the data, the total approximate expression storage means 2 stores the ID of the approximate expression, the primary The coefficient of the function and the constant term may be stored in association with each other. Note that examples of such information are merely examples, and the information stored in all approximate expression storage means 2 is not limited to these.

  FIG. 2 is an explanatory diagram showing an example of data representing the approximate expression stored in the all approximate expression storage means 2. As shown in FIG. The example shown in FIG. 2 shows a case where the data value x is approximated by a linear function x = at + b of the data generation time t. In the example shown in FIG. 2, the total approximate expression storage unit 2 stores, as each approximate expression, a set in which ID (T20), a linear function coefficient (T21), and a constant term (T22) are associated with each other. ing. Hereinafter, the coefficient (T21) of the linear function is also simply referred to as a linear coefficient.

  In FIG. 2, ID (T20) is an identifier for identifying a linear function for approximating past data values. The linear coefficient (T21) is a linear coefficient (the slope value in the linear function) related to the time t in the linear function x = at + b. The constant term (T22) is a constant term of the linear function (intercept value in the linear function).

  Specifically, the representative approximate expression management unit 1 is realized by a CPU of an information processing device that operates according to a program and a storage device such as a magnetic disk device or an optical disk device. The representative approximate expression management unit 1 has a function of selecting an approximate expression within a predetermined number from all approximate expressions stored in the all approximate expression storage means 2. The representative approximate expression management unit 1 has a function of outputting the selected approximate expression to the representative approximate expression storage means 3 and storing the approximate expression in the representative approximate expression storage means 3. The approximate expression selected by the representative approximate expression management unit 1 is hereinafter referred to as a representative approximate expression.

  FIG. 3 is a block diagram illustrating a configuration example of the representative approximate expression management unit 1. In this embodiment, the representative approximate expression management unit 1 includes a range storage means 12, a representative approximate expression first limiting means 11, and a representative approximate expression second limiting means 13, as shown in FIG.

  The value range storage unit 12 stores a range (value range) that the data value input by the data input unit 91 can take. For example, when the data to be summarized by the data summarization apparatus is the CPU usage rate [%], the value of the input data is a real number between 0 and 100 (ie, [0,100] = {x | 0 ≦ x ≦ 100 }). In this case, the value range storage means 12 stores 0 and 100 which are the minimum value and the maximum value that the input data can take. The example of such information is an example, and the information stored in the range storage unit 12 is not limited to this.

FIG. 4 is an explanatory diagram illustrating an example of data representing a range stored by the range storage unit 12. As illustrated in FIG. In the example shown in FIG. 4, the case where the value range of the data is [0, 10] is shown, and the range storage means 12 stores the maximum value T121 and the minimum value T122 of the range.

  The representative approximation formula first limiting means 11 is one in which the value at the current time belongs to the range as a candidate for the representative approximation formula from all the approximation formulas representing the past data stored in the all approximation formula storage means 2 And a function of extracting only the representative approximate expression second limiting means 13.

The representative approximate expression first limiting means 11 receives (inputs) all approximate expressions f _i (i = 1, 2,..., N) stored in the all approximate expression storage means 2 as inputs. Also, the representative approximate expression first limiting means 11 obtains an approximate value f _i (t _now ) at the current time t _now using each input approximate expression f _i . Then, the representative approximate expression first limiting means 11, of the approximate expression total approximate equation storing means 2 stores the value at the current time t _now obtained f _i (t _now) only f _i that belong to range V Is output to the representative approximation formula second limiting means 13.

  The representative approximate expression first limiting means 11 reads the approximate expression from the entire approximate expression storage means 2 at predetermined intervals and performs a process of selecting an approximate expression whose approximate value at the current time belongs to the range. May be. Alternatively, the representative approximate expression first limiting means 11 inputs an activation instruction at an arbitrary timing at which the representative approximate expression is desired to be updated according to an operation of the user, etc., reads the approximate expression from the all approximate expression storage means 2, and reads the current time You may perform the process which selects the approximate expression in which the approximate value in belongs to a range.

  The representative approximate expression second limiting means 13 has a function of receiving (inputting) the approximate expression output by the representative approximate expression first limiting means 11 as an input and selecting an approximate expression equal to or less than a predetermined number from the input. . The representative approximate expression second limiting means 13 has a function of outputting the selected (selected) approximate expression to the representative approximate expression storing means 3 and storing the approximate expression in the representative approximate expression storing means 3.

  For example, the representative approximation formula second limiting means 13 selects, for example, a predetermined number of approximation formulas from the approximate approximation formula first limit means 11 in descending order of IDs of the approximate formulas (that is, in descending order of ID values). Select (select). If the approximate expression IDs are assigned in ascending order, selecting the descending order of the IDs, the representative approximate expression second limiting means 13 selects a predetermined number of approximate expressions in the order of the newest determined approximate expressions. Can be output. Note that such a method of selecting an approximate expression is an example, and the method by which the representative approximate expression second limiting means 13 selects an approximate expression is not limited to such a method. For example, conversely, when the approximate expression ID is assigned in descending order, the representative approximate expression second limiting means 13 may select the approximate expression in ascending order of the ID.

  Specifically, the representative approximate expression storage unit 3 is realized by a storage device such as a magnetic disk device or an optical disk device. The representative approximate expression storage means 3 stores a predetermined maximum number N of representative approximate expressions selected by the representative approximate expression management unit 1. The representative approximate expression storage means 3 may store data in the same data format as the data stored in the all approximate expression storage means 2, for example. For example, when the approximate expression is represented by a linear function at time t, the representative approximate expression storage unit 3 may store the ID of the approximate expression, the coefficient of the linear function, and the constant term in association with each other. Further, the representative approximate expression storage unit 3 may store only the ID of the approximate expression, for example. Note that examples of such information are merely examples, and the information stored in the representative approximate expression storage unit 3 is not limited to these.

  FIG. 5 is an explanatory diagram showing an example of data representing the representative approximate expression stored in the representative approximate expression storage means 3. The example shown in FIG. 5 shows a case where the data value x is approximated by a linear function x = at + b of time. In the example shown in FIG. 5, the representative approximate expression storage unit 3 stores, as each approximate expression, a set in which an ID (T30), a coefficient of a linear function (T31), and a constant term (T32) are associated with each other. ing.

  Specifically, the data input means 91 is realized by a CPU and an input interface unit of an information processing apparatus that operates according to a program. The data input means 91 has a function of acquiring (inputting) data from a data generation source that sequentially generates data over time. The data generation source varies depending on the type of data. For example, when the CPU usage rate is used as input data, a unit that monitors the CPU usage rate may be used as a data generation source. For example, when the number of accesses to a web page is used as input data, a web server may be used as a data generation source. The data input means 91 inputs data including at least a data value and an occurrence time of the data value as individual data from the data generation source.

  Specifically, the approximating means 92 is realized by a CPU of an information processing device that operates according to a program and a storage device such as a magnetic disk device or an optical disk device. The approximating means 92 has a function of receiving (inputting) the data acquired (input) by the data input means 91 and the approximate expression stored in the representative approximate expression storage means 3 as inputs. Further, the approximation means 92 approximates the input data by the approximate expression if there is an approximate expression input from the representative approximate expression storage means 3 that can represent the data input from the data input means 91 with a sufficiently small error. It has a function. Further, if there is no sufficiently small error approximate expression, the approximating means 92 creates a new approximate expression and outputs it to the all approximate expression storing means 2, and the newly created approximate expression is stored in the all approximate expression storing means 2 The function to memorize is provided. The approximating unit 92 operates every time data is input from the data input unit 91, and executes a process of approximating input data using an approximate expression and a process of generating an approximate expression.

  FIG. 6 is a block diagram showing an example of the configuration of the approximating means 92. As shown in FIG. In the example shown in FIG. 6, the approximating unit 92 includes an approximate formula calculating unit 921, an output unit 922, an undetermined data storage unit 923, and a new approximate formula generating unit 924.

  The approximate expression calculation unit 921 has a function of calculating a value (approximate value) corresponding to the generation time of the data value input from the data input unit 91 with respect to the approximate expression stored in the representative approximate expression storage unit 3. That is, the approximate expression calculation unit 921 substitutes the value of the occurrence time into the approximate expression, and obtains an approximate value when the input data value is approximated using the approximate expression. The approximate expression calculation unit 921 has a function of selecting an approximate expression suitable for calculating the approximate value of the input data based on the value for the obtained approximate expression and the data value input from the data input unit 91. That is, for example, the approximate expression calculation unit 921 calculates an error between the approximate value obtained using the approximate expression and the actual data input from the data input unit 91, and the error is within a predetermined allowable value and is the most error. Is selected as the optimum approximation formula. The approximate expression calculation unit 921 has a function of outputting the selected approximate expression and the generation time of the data value to the output unit 922.

  On the other hand, if there is no suitable approximation formula (for example, if there is no approximation formula with an error within a predetermined allowable value), the approximate formula calculation means 921 sends the data value input from the data input means 91 to the undefined data storage means 923. Is output and stored as it is.

  For example, the approximate expression calculation means 921 inputs each approximate value calculated by substituting the generation time of the data input from the data input means 91 for each approximate expression stored in the representative approximate expression storage means 3. Compare the actual data value of the data. Then, the approximate expression calculation unit 921 identifies one approximate expression according to a predetermined condition among approximate expressions that satisfy a predetermined accuracy.

  Note that the method of specifying the approximate expression is arbitrary. For example, the approximate expression calculation unit 921 specifies an approximate expression having the best accuracy among the approximate expressions satisfying a predetermined accuracy, and the specified approximate expression is It may be determined that the input data value is used to approximate. Alternatively, when the approximate expression calculation means 921 is selected as an approximate expression representing the input data, for example, the increase in the total storage capacity used in the storage means (for example, a magnetic disk device or an optical disk device) of the data summarizing device is minimized. It is also possible to determine that the approximate expression that becomes is used and to use the specified approximate expression to approximate the input data value.

  Alternatively, the approximate expression calculation unit 921 determines, for example, that an approximate expression that minimizes the amount of calculation of the value when substituting time is specified, and that the specified approximate expression is used to approximate the input data value. You can do it. Further, the approximate expression calculation unit 921 may determine, for example, that an approximate expression with the latest time at which the approximate expression is generated is specified and that the specified approximate expression is used to approximate the input data value.

  Note that the approximate expression specifying method described above is merely an example, and the method by which the approximate expression calculating unit 921 specifies the approximate expression is not limited thereto.

  On the other hand, the approximate expression calculation unit 921 may not be able to specify one approximate expression from among approximate expressions that satisfy a predetermined accuracy. That is, the approximate expression calculation unit 921 cannot specify one approximate expression when there is no approximate expression satisfying the predetermined accuracy in the representative approximate expression storage unit 3. In this case, the approximate expression calculation means 921 does not select the approximate expression, but outputs and stores the new data input from the data input means 91 to the unconfirmed data storage means 923.

Next, an example of the approximate expression specifying process by the approximate expression calculating unit 921 will be described in detail. FIG. 7 is an explanatory diagram illustrating an example of selecting an approximate expression that satisfies a predetermined accuracy constraint. In the example shown in FIG. 7, the generation time of new data input to the approximate expression calculation means 921 is t _i and the data value is x _i . In the example shown in FIG. 7, it is assumed that new data P1022 of value x _i is generated at time t _i and is input from the data input means 91 to the approximate expression calculation means 921. The representative approximate expression storage means 3 stores approximate expressions x = f ₀ (t), x = f ₁ (t), x = f ₂ (t), and x = f ₃ (t). Shall. Further, the accuracy constraint required for the approximate expression is that the difference between the approximate value and the actual data is less than ε.

The approximate expression calculation means 921 substitutes time t = t _i for each approximate expression, and approximate values X ₁₀₁₀ = f ₀ (t _i ), X ₁₀₁₁ = f ₁ (t _i ), X ₁₀₁₂ = Calculate f ₂ (t _i ) and X ₁₀₁₃ = f ₃ (t _i ). In the example shown in FIG. 7, the approximate values X ₁₀₁₂ = f ₂ (t _i ) and X ₁₀₁₃ = f ₃ (t _i ) calculated using the approximate expressions f ₂ and f ₃ satisfy the accuracy constraint value ε. Yes. Therefore, the approximate expression calculation unit 921 identifies one approximate expression according to a predetermined condition from the approximate expressions f ₂ and f _{3 that} satisfy the accuracy constraint.

For example, when one approximate expression is specified according to the condition that the accuracy of the approximate expression is specified, the approximate expression calculation unit 921 uses | x _i -f ₃ (t _{i in the} example shown in FIG. ) | <| X _i −f ₂ (t _i ) | <ε, the approximate expression f ₃ is identified and output to the output unit 922.

  The indeterminate data storage unit 923 is a storage device that stores data for which an approximate expression for obtaining an approximate value of a data value has not yet been specified. Even if the approximate value is calculated using any of the existing approximate expressions stored in the representative approximate expression storage means 3, the data that is determined to have a large difference between the actual data value and the approximate value is an undetermined data storage means It will be remembered in 923. In the present embodiment, since the data can be regarded as a point having the occurrence time and the data value as coordinates, the point on the coordinate where the approximate expression of the data is located is also expressed as “undefined point”. To do.

  An approximate expression for obtaining an approximate value of the data value using the occurrence time as a variable, which is newly generated when a representative approximate expression that approximates the input data is not found, is a plurality of unconfirmed data stored in the unconfirmed data storage means 923. Derived based on points. In the present embodiment, the undetermined data storage unit 923 stores data corresponding to the undetermined points until the necessary number of pieces of data for determining the approximate expression are prepared. Note that the number of data necessary to determine the approximate expression depends on the type of approximate expression (whether it is a linear expression, a quadratic expression, or an expression using a trigonometric function), an approximation, etc. Depends on the formula decision algorithm. It is assumed that the type of approximate expression and the approximate expression determination algorithm are determined in advance, and the number of data required to determine the approximate expression is determined in advance according to the type of approximate expression and the approximate expression determination algorithm.

  The new approximate expression generation unit 924 has the data necessary for determining the approximate expression for the uncertain points stored in the unconfirmed data storage unit 923 when new data is input. Then, a function (approximate expression) for generating a data value using the occurrence time as a variable is provided based on the data.

  For example, it is assumed that the number of data required for generating the approximate expression is k, and k−1 uncertain points are stored in the unconfirmed data storage unit 923. At this time, when one piece of data is newly input and stored in the unconfirmed data storage unit 923, and k pieces of data are prepared, the new approximate expression generation unit 924 stores the k data stored in the unconfirmed data storage unit 923. A new approximate expression is generated based on the pieces of data. Further, the new approximate expression generation unit 924 stores the generated approximate expression in the all approximate expression storage unit 2. Also, the new approximate expression generation unit 924 outputs, to the output unit 922, a set in which the generation time of the data and the newly generated approximate expression are associated with each other for the k pieces of data. In other words, the new approximate expression generation unit 924 outputs k sets of sets of data generation times and newly generated approximate expressions.

  The output unit 922 outputs a set of one or more approximate formulas input from the approximate formula calculation unit 921 or the new approximate formula generation unit 924 and the data generation time. In this case, for example, when the output unit 922 is realized by a display device such as a display device, the output unit 922 displays a set of the approximate expression and the data generation time. In addition, for example, the output unit 922 may output a file including a set of an approximate expression and data generation time.

  Note that the configuration example of the approximating unit 92 described above is an example, and the configuration of the approximating unit 92 is not limited to this.

  In the present embodiment, the storage device of the data summarization device stores various programs for executing processing for summarizing sequentially input data. For example, the storage device of the data summarization apparatus can be used as a representative approximate expression that is a candidate for approximating newly generated data in a computer including all approximate expression storage means for storing a plurality of approximate expressions for approximating data. The approximate value obtained by using each approximate expression based on the current time from the approximate expressions stored in the all approximate expression storage means is included in the range of values that the data can take. And a data summarizing program for executing a process of selecting an approximate expression within a predetermined number and a process of storing the selected representative approximate expression in the representative approximate expression storage means.

Next, the operation of the data summarization apparatus will be described. First, processing for selecting a representative approximate expression will be described. FIG. 8 is a flowchart showing a processing example of representative approximate expression selection processing executed by the data summarization apparatus. First, the representative approximate expression first limiting means 11 is the only one of all approximate expressions stored in the all approximate expression storage means 2 that belongs to the value range in which the value at the current time is stored in the value range storage means 12. Is output to the representative approximate expression second limiting means 13 (step S11). In this case, it is assumed that the representative approximate expression first limiting means 11 selects and outputs the approximate expressions f ₁ , f ₂ ,..., F _n .

  Next, the representative approximate expression second limiting means 13 compares the number n of input approximate expressions with a predetermined number N (step S12). As a result of the comparison, if it is determined that n ≦ N (Yes in step S12), the representative approximate expression second limiting means 13 stores all the approximate expressions input in step S11, whereby the representative approximate expression The stored contents of the storage means 3 are updated (step S14). On the other hand, if it is determined that n> N (No in step S12), the representative approximate expression second limiting means 13 removes nN approximate expressions from the approximate expression input in step S11, and N N Is selected (step S13). Then, the representative approximate expression second limiting means 13 updates the stored contents of the representative approximate expression storage means 3 by storing the remaining N approximate expressions (step S14).

Next, the operation of the approximate expression selection process executed by the representative approximate expression first limiting means 11 in step S11 will be described. Representative approximate expression first limiting means 11, for all of the approximation formula f _i which is stored in the total approximate equation storing means 2 performs the following process. Figure 9 is a flow chart showing an example of approximate expression selection processing representative approximate expression first limiting means 11 for the respective approximation formula f _i which is stored in the total approximate equation storing means 2 is performed in step S11.

First, the representative approximate expression first limiting means 11 obtains the value f _i (t _now ) of the approximate expression f _{i at} the current time t _now (step S111). Next, the representative approximation formula first limiting means 11 determines whether or not f _i (t _now ) ∈V for the obtained approximate value f _i (t _now ) and the range V stored in the range storage unit 12. Is determined (step S112). Results of the evaluation, f _i (t _now) in the case of determining that a ∈ V (Yes in step S112), the representative approximate expression first limiting means 11, the second limiting means 13 representative approximation formula the approximation formula f _i (Step S113). On the other hand, if it is determined that f _i (t _now ) ∈V does not hold (No in step S112), the representative approximation formula first limiting means 11 does not proceed to step S113, and the approximate formula f _i is represented as a representative approximation. Control is performed so as not to output to the second limiting means 13.

  According to the above processing, an approximate expression satisfying a predetermined accuracy condition and satisfying a predetermined selection condition is selected from the approximate expressions stored in the all approximate expression storage means 2, and the representative approximate expression storage means 3 is selected as a representative approximate expression. Remembered. After that, the data summarization apparatus performs input data summarization processing based on the representative approximate expression selected and stored in the representative approximate expression storage means 3. It should be noted that the accuracy of the representative approximate expression stored in the representative approximate expression storage means 3 is maintained by the data summarizing apparatus repeatedly executing the process of selecting the representative approximate expression shown in FIGS. 8 and 9 at predetermined intervals or at predetermined timings. Be drunk.

  As described above, according to the present embodiment, instead of performing summarization processing using all approximate expressions generated in the past, even if the number of approximate expressions continues to increase, the representative approximate expression storage unit 2 Summarization processing is performed using only the stored representative approximation formulas within a predetermined number. Therefore, data summarization can be continued even if the period of data summarization becomes longer. Even in such a configuration, the accuracy of the data summarization can be prevented from being lowered more than necessary because the summarization process can be performed using the representative approximate expression selected based on the predetermined condition while satisfying the predetermined accuracy condition. it can. Therefore, when the value of data changes suddenly, data summarization can be continued even if the period for which data summarization is performed becomes longer.

(Second embodiment)
Next, a second embodiment of the present invention will be described with reference to the drawings. FIG. 10 is a block diagram illustrating a configuration example of the data summarizing device according to the second embodiment. In the present embodiment, components similar to those in the first embodiment are denoted by the same reference numerals as those in FIG. 1, and description thereof is omitted. As shown in FIG. 10, in this embodiment, the data summarizing device includes a range storage means 12, a representative approximate expression first limiting means 11, a representative approximate expression second limiting means 13, and a new approximate expression addition managing means 21. A total approximate expression storage means 2, a representative approximate expression storage means 3, and an approximate expression update means 95. Therefore, in this embodiment, the data summarizing apparatus includes approximate expression updating means 95 in addition to the components shown in the first embodiment. Further, the representative approximate expression management unit 1 of the data summarizing apparatus includes a new approximate expression addition management means 21 in addition to the components shown in the first embodiment.

  Specifically, the approximate expression update unit 95 is realized by a CPU of an information processing apparatus that operates according to a program. The approximate expression update unit 95 has a function of updating an approximate expression as needed by generating a new approximate expression and storing it in the all approximate expression storage means 2. The approximate expression update unit 95 has a function of outputting a new approximate expression to the new approximate expression addition management unit 21. Note that the approximate expression update unit 95 may perform a process of storing the newly input approximate expression in the all approximate expression storage unit 2, for example. Further, the approximate expression update unit 95 may be realized by the new approximate expression generation unit 924 shown in the first embodiment, for example.

  Specifically, the new approximate expression addition management means 21 is realized by a CPU of an information processing apparatus that operates according to a program. The new approximate expression addition management means 21 has a function of receiving (inputting) a new approximate expression to be added to the all approximate expression storage means 2 from the approximate expression update means 95 as an input. Further, when the new approximate expression addition managing means 21 determines that the representative approximate expression first limiting means 11 should be activated, a new representative approximate expression candidate including the input approximate expression is assigned to the representative approximate expression first limiting means 11. A function of starting the representative approximate expression first limiting means 11 by outputting is provided.

  The activation of the representative approximate expression first limiting means 11 by the new approximate expression addition managing means 21 may be performed every time input from the approximate expression updating means 95, or a certain number of inputs are received from the approximate expression updating means 95. You can go every time. Alternatively, it may be performed at every elapse of a certain time or more after the new approximate expression addition management means 21 is activated last, or may be performed by a combination of these plural activation conditions. Note that such a method of starting the representative approximate expression first limiting means 11 is an example, and the method of starting the representative approximate expression first limiting means 11 is not limited to these.

Next, the operation of the data summarization apparatus will be described. FIG. 11 is a flowchart illustrating a processing example of representative approximate expression selection processing executed by the data summarizing apparatus according to the second embodiment. First, when a new approximate expression addition management means 21 inputs a new approximate expression from the approximate expression update means 95 (step S21), it determines whether or not the representative approximate expression should be updated (step S22). In this case, the new approximate expression input from the approximate expression update means 95 is defined as _fNEW . If it is determined that the representative approximate expression should be updated (Yes in step S22), the new approximate expression addition managing means 21 is after the last update of the representative approximate expression (that is, during the last representative approximate expression selection process). the added approximate expression to extract from all the approximate equation storing means 2, and outputs with a new approximation formula f _nEW to the first limiting means 11 representative approximate equation (step S23).

For example, when the approximate expressions g ₁ , g ₂ ,..., G _m are input from the approximate expression update means 95 after the previous update process of the representative approximate expressions stored in the representative approximate expression storage means 3 The new approximate expression addition management means 21 reads the approximate expressions g ₁ , g ₂ ,..., G _m from the all approximate expression storage means 2. Then, the new approximate expression addition management means 21 outputs the new approximate expression f _NEW to the representative approximate expression first limiting means 11 together with the read approximate expressions g ₁ , g ₂ ,..., G _m .

  On the other hand, when it is determined that the representative approximate expression should not be updated (No in step S22), the data summarizing apparatus ends the process without updating the representative approximate expression.

  Next, the representative approximation formula first limiting means 11 stores the value at the current time among the current approximation formula stored in the representative approximation formula storage means 3 and the approximation formula input from the new approximation formula addition management means 21 in the range storage. Only those belonging to the value range stored in the means 12 are taken out and output to the representative approximation formula second limiting means 13 (step S24). Then, the data summarizing apparatus performs processing for updating the representative approximate value according to processing similar to steps S12 to S14 described in the first embodiment.

  As described above, according to the present embodiment, the representative approximate expression selection processing by the representative approximate expression first limiting means 11 and the representative approximate expression second limiting means 13 depends on the timing of determining a new approximate expression. Done. Therefore, even when the data generation timing is irregular and the timing of determining the new approximate expression is unpredictable, the representative approximate expression selection process can be performed according to the determination of the new approximate expression. The approximate expression can be a representative approximate expression.

(Third embodiment)
Next, a third embodiment of the present invention will be described with reference to the drawings. FIG. 12 is a block diagram illustrating a configuration example of the data summarizing device according to the third embodiment. In the present embodiment, components similar to those in the first embodiment are denoted by the same reference numerals as those in FIG. 1, and description thereof is omitted. As shown in FIG. 12, in the present embodiment, the data summarizing device includes a range storage unit 12, a representative approximate expression first limiting unit 11, a representative approximate expression second limiting unit 13, and a total approximate expression storing unit 2. The representative approximate expression storage means 3, the representative approximate expression resource monitoring means 31, and the representative approximate expression management control means 32 are included. Therefore, in this embodiment, the representative approximate expression management unit 1 of the data summarizing apparatus includes a representative approximate expression resource monitoring unit 31, a representative approximate expression management control unit 32, in addition to the components shown in the first embodiment. including.

  Specifically, the representative approximate expression resource monitoring unit 31 is realized by a CPU of an information processing apparatus that operates according to a program. The representative approximate expression resource monitoring means 31 has a function of monitoring at least the amount of resources that can be used in the area storing the representative approximate expression and outputting the monitoring result to the representative approximate expression management control means 32. In other words, the representative approximate expression resource monitoring unit 31 uses the amount of resources that can be used in the representative approximate expression storage unit 3 (for example, storage that can be allocated as the representative approximate expression storage unit 3 in the storage area of the storage device included in the information processing device). Monitor the capacity of the area. For example, if the representative approximate expression storage means 3 is realized on the same storage device as the other storage means, the representative approximate expression resource monitoring means 31 may monitor the amount of resources that can be used in the storage apparatus. .

  For example, the representative approximate expression resource monitoring unit 31 monitors the free space of the physical memory that realizes the representative approximate expression storage unit 3 as an example of the amount of resources that can be used as the representative approximate expression storage unit 3. The example of the resource monitored by the representative approximate resource monitoring unit 31 is merely an example, and the resource monitored by the representative approximate resource monitoring unit 31 is not limited to this. For example, when the representative approximate expression storage unit 3 is realized by a disk device (magnetic disk device or optical disk device), the representative approximate expression resource monitoring unit 31 may monitor the unused rate of the disk device. Further, the representative approximate resource monitoring unit 31 may monitor, for example, the resource amount that is already in use (for example, the used capacity of the memory) instead of the available resource amount. Further, the representative approximate expression resource monitoring unit 31 may monitor, for example, a resource (for example, CPU time) used for approximation evaluation using the representative approximate expression as the monitoring target resource amount.

  Further, the representative approximate expression resource monitoring unit 31 may monitor the representative approximate expression storage unit 3 at regular intervals, for example, as the resource monitoring timing. Alternatively, the representative approximate resource monitoring unit 31 may monitor resources when an instruction to monitor at an arbitrary timing is input in accordance with, for example, a user operation.

  Specifically, the representative approximate expression management control means 32 is realized by a CPU of an information processing apparatus that operates according to a program. The representative approximate expression management control means 32 has a function of determining the maximum number N of representative approximate expressions stored in the representative approximate expression storage means 3 in accordance with the monitoring result input from the representative approximate expression resource monitoring means 31. Further, the representative approximate expression management control means 32 has a function of starting the representative approximate expression first limiting means 11 and updating the representative approximate expression when the maximum number N changes.

For example, the maximum number of candidates N _small and N _large (N _small <N _large ) stored in the representative approximate expression storage unit 3 is determined in advance, and the representative approximate expression management control unit 32 sets the representative approximate expression resource It is determined whether or not the available resource amount input as a monitoring result from the monitoring unit 31 is equal to or greater than a threshold value. If it is equal to or greater than the threshold value, the representative approximate expression management control means 32 selects N _large as the maximum number and controls the representative approximate expression storage means 3 to store N _large representative approximate expressions. On the other hand, if it is less than the threshold value, the representative approximate expression management control means 32 may select N _small as the maximum number and control to store N _small representative approximate expressions. That is, the maximum number of representative approximate expressions may be increased if there is a margin in resources, and the maximum number of representative approximation expressions may be decreased if there is not enough resources.

Alternatively, for example, N ₁ , N ₂ ,..., N _k (N ₁ <N ₂ <... <N _k ) are determined in advance as the maximum number of representative approximate expressions stored in the representative approximate expression storage means 3. The representative approximate expression management control means 32 receives the resource amount R that can be input from the representative approximate expression resource monitoring means 31 as a monitoring result and k−1 threshold values T ₁ , T ₂ ,..., T _k−1 ( T ₁ <T ₂ <... <T _k-1 ). Then, the representative approximate expression management control means 32 determines that the maximum number N of representative approximate expressions stored in the representative approximate expression storage means 3 is N = N ₁ when it is determined that R ≦ T _1, and T _i− If it is determined that ₁ <R ≦ T _i , N = N _i (i = 1, 2, .., k−1) is determined, and if T _k−1 <R is determined, N = N _k May be determined. Alternatively, for example, the representative approximate expression management control means 32 determines that N = N ₁ when the maximum number N of representative approximate expressions is determined as R <T ₁ and T _i−1 ≦ R <T _i If determined, N = N _i (i = 1, 2, .., k−1) may be determined, and if T _k−1 ≦ R, N = N _k may be determined. . Alternatively, the representative approximate expression management control means 32 may determine that the maximum number N of representative approximate expressions is N = N (R) by a bounded monotonically increasing function N (R) with respect to the amount of available free resources. Good. Note that the method for determining the maximum number N of such representative approximate expressions is an example, and the method for determining the maximum number N is not limited to this.

  As described above, according to the present embodiment, the maximum number N of representative approximate expressions is determined according to the amount of resources for storing the representative approximate expressions. Therefore, if there is a margin in resources, the data values can be approximated with higher accuracy by using many representative approximation formulas. On the other hand, if there are not enough resources, the number of representative approximate expressions can be reduced, and data summarization can be performed using the minimum necessary representative approximate expressions.

  Note that each means described in each of the above embodiments may be realized by using a CPU, a memory, and a disk device on one computer. Alternatively, each means shown in each of the above embodiments may be realized using a plurality of computers connected to each other via a network.

  Next, the minimum configuration of the data summarizing apparatus according to the present invention will be described. FIG. 13 is a block diagram showing a minimum configuration example of the data summarizing apparatus. As shown in FIG. 13, the data summarizing apparatus includes a representative approximate expression management unit 1, a total approximate expression storage means 2, and a representative approximate expression storage means 3 as minimum components.

  In the data summarizing apparatus with the minimum configuration shown in FIG. 13, the representative approximate expression storage means 3 stores a plurality of approximate expressions for approximating data. Further, the representative approximate expression management unit 1 selects each approximation based on the current time from the approximate expressions stored in the all approximate expression storage means 2 as representative approximate expressions that are candidates for approximating newly generated data. It has a function of selecting an approximate expression that is included in the range of the range that is the range of values that the data can take as the approximate value obtained using the formula. The representative approximate expression storage unit 3 stores the representative approximate expression selected by the representative approximate expression management unit 1.

  According to the data summarizing apparatus having the minimum configuration shown in FIG. 13, data summarization can be continued even when the period of data summarization becomes longer when the value of the data suddenly changes.

  In each of the embodiments described above, the characteristic configuration of the data summarization apparatus as shown in the following (1) to (8) is shown.

  (1) The data summarization device is a data summarization device for summarizing sequentially generated data, and is realized by a total approximation expression storage means (for example, a total approximation expression storage means 2) for storing a plurality of approximation expressions for approximating data. As a representative approximate expression that is a candidate for approximating newly generated data, the approximate expression stored in the all approximate expression storage means is obtained using each approximate expression based on the current time. Representative approximate expression selection means (for example, realized by the representative approximate expression management unit 1) for selecting an approximate expression included in the range of the range where the approximate value is a range of values that the data can take, and representative approximation And a representative approximate expression storage means (for example, realized by the representative approximate expression storage means 3) for storing the representative approximate expression selected by the expression selection means.

  (2) The data summarization device is a data summarization device for summarizing sequentially generated data, and is realized by a total approximation expression storage means (for example, a total approximation expression storage means 2) for storing a plurality of approximate expressions for approximating data. As a representative approximate expression that is a candidate for approximating newly generated data, the approximate expression stored in the all approximate expression storage means is obtained using each approximate expression based on the current time. Representative approximate expression selection means (for example, representative approximate expression management unit 1) that selects an approximate expression within the range of values that are values that the data can take and that is within a predetermined number. And a representative approximate expression storage means (for example, realized by the representative approximate expression storage means 3) for storing the representative approximate expression selected by the representative approximate expression selecting means.

  (3) The data summarization device is a data summarization device for summarizing sequentially generated data, and is realized by a total approximation expression storage means (for example, a total approximation expression storage means 2) for storing a plurality of approximate expressions for approximating data. Selected from the multiple approximation formulas stored in the total approximation formula storage means as a representative approximation formula that is a candidate for approximating newly generated data. Representative approximate expression selection means (for example, realized by the representative approximate expression management unit 1) and representative approximate expression storage means for storing the representative approximate expression selected by the representative approximate expression selection means (for example, representative approximate expression storage means 3) It is characterized by the above.

  (4) In the data summarizing apparatus, the representative approximate expression selecting means includes a range storage means (for example, realized by the range storage means 12) for storing range information indicating a range of values that can be taken by the data, and an all approximate expression storage. An approximate value obtained using each approximate expression based on the current time is included in the range of values indicated by the range information stored in the range storage means from all the approximate expressions stored by the means. First approximation formula selection means (for example, realized by the representative approximation formula first limiting means 11) that selects only the approximation formula and an approximation formula selected by the first approximation formula selection means are predetermined. The second approximate expression selecting means (for example, realized by the representative approximate expression second limiting means 13) for selecting the approximate expression within the predetermined number as the representative approximate expression may be included.

  (5) In the data summarization apparatus, the representative approximation formula selection means includes a range storage means (for example, realized by the range storage means 12) for storing range information indicating a range of values that the data can take, and an approximate expression. An approximate expression update determination means for determining whether or not the representative approximate expression stored in the representative approximate expression storage means should be updated (for example, realized by the new approximate expression addition management means 21), and an approximate expression update When the determination means determines that the representative approximate expression should be updated, the representative approximate expression stored in the representative approximate expression storage means and the representative approximate expression stored in the representative approximate expression storage means are added after the past update. Only approximate expressions that are included in the range of values indicated by the range information stored in the range storage means, are approximate values obtained using the approximate formulas based on the current time. Select Of the approximate expressions selected by the first approximate expression selecting means (for example, realized by the representative approximate expression first limiting means 11) and the approximate expression selected by the first approximate expression selecting means, an approximate expression within a predetermined number. The second approximate expression selecting means (for example, realized by the representative approximate expression second limiting means 13) may be included.

  (6) In the data summarizing device, the representative approximate expression selecting means is representative approximate expression resource monitoring means (for example, realized by the representative approximate expression resource monitoring means 31) for monitoring the resource amount usable by the representative approximate expression storage means. And a representative approximate expression update control means for determining the number of representative approximate expressions to be stored in the representative approximate expression storage means and updating the representative approximate expression storage means (for example, , Realized by the representative approximate expression management control means 32).

  (7) In the data summarizing apparatus, the representative approximate expression selecting means includes a range storage means (for example, realized by the range storage means 12) for storing range information indicating a range of values that can be taken by the data, and an all approximate expression storage. An approximate value obtained using each approximate expression based on the current time is included in the range of values indicated by the range information stored in the range storage means from all the approximate expressions stored by the means. A first approximate expression selecting means (for example, realized by the representative approximate expression first limiting means 11) for selecting only the approximate expression and a representative approximate expression based on the approximate expression selected by the first approximate expression selecting means Second approximate expression selection means (for example, realized by the representative approximate expression second limiting means 13) to select, the second approximate expression selection means is the approximate expression selected by the first approximate expression selection means Whether or not is within a predetermined number If it is determined that the number is within the predetermined number, the approximate expression selected by the first approximate expression selecting unit is determined as the representative approximate expression as it is, and if it is determined that the number is not within the predetermined number The approximate expression selected by the first approximate expression selecting means may be configured to select an approximate expression within a predetermined number as the representative approximate expression.

  (8) The data summarizing apparatus includes approximation means (for example, realized by the approximation means 92) for approximating data using each representative approximation expression stored in the representative approximation expression storage means. Based on the error between the approximate value calculated using each representative approximate expression stored in the storage means and the value of the actual data, the approximation used for data approximation from the representative approximate expressions stored in the representative approximate expression storage means When there is no approximate expression that can be used to approximate the data in the approximate expression determining means (for example, realized by the approximate expression calculating means 921) for determining the expression and each representative approximate expression stored in the representative approximate expression storing means In addition, an approximate expression generation unit (for example, realized by the new approximate expression generation unit 924) that newly generates an approximate expression may be included.

  The present invention can be suitably applied to the use of a data summarization apparatus that summarizes data using an approximation process using an approximate expression.

1 Representative approximation formula management department
2 All approximation formula storage means
3 Representative approximation formula storage means
11 Representative approximation formula 1st limiting means
12 Range storage means
13 Representative approximation formula second limiting means
21 New approximation formula addition management means
31 Representative approximate resource monitoring means
32 Representative approximate expression management control means
91 Data input means
92 Approximate means
95 Approximation formula update method
921 Approximation formula calculation means
922 Output means
923 Indeterminate data storage means
924 New approximate expression generation means

Claims (10)

A data summarization device for summarizing sequentially generated data,
All approximate expression storage means for storing a plurality of approximate expressions for approximating data;
As a representative approximate expression that becomes a candidate for approximating newly generated data, an approximate value obtained by using each approximate expression based on the current time is selected from the approximate expressions stored in the all approximate expression storage means. Representative approximate expression selecting means for selecting an approximate expression included in a range of a value range that can be taken by data;
A data summarizing apparatus comprising: a representative approximate expression storing means for storing the representative approximate expression selected by the representative approximate expression selecting means. A data summarization device for summarizing sequentially generated data,
All approximate expression storage means for storing a plurality of approximate expressions for approximating data;
As a representative approximate expression that becomes a candidate for approximating newly generated data, an approximate value obtained by using each approximate expression based on the current time is selected from the approximate expressions stored in the all approximate expression storage means. Representative approximate expression selection means for selecting an approximate expression that is included in a range of values that can be taken by the data and that is within a predetermined number,
A data summarizing apparatus comprising: a representative approximate expression storing means for storing the representative approximate expression selected by the representative approximate expression selecting means. A data summarization device for summarizing sequentially generated data,
All approximate expression storage means for storing a plurality of approximate expressions for approximating data;
Representative approximate expression for selecting an approximate expression within a predetermined number from a plurality of approximate expressions stored in the all approximate expression storage means as a representative approximate expression to be a candidate for approximating newly generated data A selection means;
A data summarizing apparatus comprising: a representative approximate expression storing means for storing the representative approximate expression selected by the representative approximate expression selecting means. The representative approximate expression selection means includes:
Range storage means for storing range information indicating the range of values that the data can take;
A value indicated by the range information stored in the range storage means is an approximate value obtained by using each approximation formula based on the current time among all the approximate formulas stored in the total approximation formula storage means First approximation formula selection means for selecting only the approximation formula included in the range of
The second approximate expression selecting means for selecting, as the representative approximate expression, an approximate expression within a predetermined number from the approximate expressions selected by the first approximate expression selecting means. 4. The data summarization device according to any one of 3. The representative approximate expression selection means includes:
Range storage means for storing range information indicating the range of values that the data can take;
Approximate expression update determination means for determining whether or not the representative approximate expression stored in the representative approximate expression storage means should be updated when an approximate expression is added;
The representative approximate expression stored in the representative approximate expression storage means and the representative approximate expression stored in the representative approximate expression storage means when the approximate expression update determination means determines that the representative approximate expression should be updated. The approximate value obtained by using each approximate expression based on the current time from the approximate expressions added after the time is updated in the past is indicated by the range information stored in the range storage means. A first approximation formula selection means for selecting only the approximation formula included in the value range;
The second approximate expression selecting means for selecting, as the representative approximate expression, an approximate expression within a predetermined number from the approximate expressions selected by the first approximate expression selecting means. 4. The data summarization device according to any one of 3. The representative approximate expression selection means includes:
Representative approximate expression resource monitoring means for monitoring the amount of resources that can be used by the representative approximate expression storage means;
Based on the amount of resources that can be used by the representative approximate expression storage means, the representative approximate expression update control means for determining the number of the representative approximate expressions to be stored in the representative approximate expression storage means and updating the representative approximate expression storage means The data summarization device according to any one of claims 1 to 4. The representative approximate expression selection means includes:
Range storage means for storing range information indicating the range of values that the data can take;
A value indicated by the range information stored in the range storage means is an approximate value obtained by using each approximation formula based on the current time among all the approximate formulas stored in the total approximation formula storage means First approximation formula selection means for selecting only the approximation formula included in the range of
A second approximate expression selecting means for selecting the representative approximate expression based on the approximate expression selected by the first approximate expression selecting means,
The second approximate expression selecting means is
Determining whether the approximate expression selected by the first approximate expression selecting means is within a predetermined number,
When it is determined that the number is within a predetermined number, the approximate expression selected by the first approximate expression selection unit is determined as the representative approximate expression as it is,
If it is determined that the number is not less than a predetermined number, an approximate expression within a predetermined number is selected as the representative approximate expression from the approximate expressions selected by the first approximate expression selection unit. The data summarization device according to any one of claims 1 to 6. Approximating means for approximating data using each representative approximate expression stored in the representative approximate expression storing means;
The approximation means includes
Based on the error between the approximate value calculated using each representative approximate expression stored in the representative approximate expression storage means and the value of the actual data, the representative approximate expression stored in the representative approximate expression storage means Approximation expression determining means for determining an approximation expression used for data approximation;
The approximate expression generation means for newly generating an approximate expression when there is no approximate expression usable for data approximation in the representative approximate expressions stored in the representative approximate expression storage means. Item 8. The data summarization device according to any one of items 7 to 9. A data summarization method for summarizing sequentially generated data,
A plurality of approximation formulas for approximating data are stored in all approximation formula storage means,
As a representative approximate expression that becomes a candidate for approximating newly generated data, an approximate value obtained by using each approximate expression based on the current time is selected from the approximate expressions stored in the all approximate expression storage means. , Select an approximate expression that is included in the range of values that can be taken by the data, and within a predetermined number,
A data summarizing method, wherein the selected representative approximate expression is stored in a representative approximate expression storage means. A data summarization program for summarizing sequentially generated data,
In a computer provided with all approximate expression storage means for storing a plurality of approximate expressions for approximating data,
As a representative approximate expression that becomes a candidate for approximating newly generated data, an approximate value obtained by using each approximate expression based on the current time is selected from the approximate expressions stored in the all approximate expression storage means. A process of selecting an approximate expression that is included in a range of values that can be taken by the data and that is within a predetermined number;
A data summarization program for executing the process of storing the selected representative approximate expression in the representative approximate expression storage means.

Images