CN107316200B - Method and device for analyzing user behavior period - Google Patents

Abstract

The application provides a method and a device for analyzing a user behavior cycle, wherein the method comprises the following steps: acquiring a time interval sequence of user behaviors; searching for a periodic time interval subsequence in which the fluctuation of adjacent elements in the time interval sequence meets a predetermined condition; calculating scores for all of the periodic time interval subsequences; and carrying out grading evaluation on the user according to the score. The method searches the subsequences presented in a periodic form in the sequence by judging whether the fluctuation degree of the user behavior time interval forming the sequence exceeds a threshold value, and scores all the subsequences, so that the analysis result of the user can be obtained.

Description

A method and device for analyzing user behavior cycle

technical field

The present application relates to the technical field of user behavior analysis, and in particular, to a method and apparatus for analyzing a user behavior cycle.

Background technique

Many existing technologies try to analyze the user's transaction behavior to obtain the user's potential transaction behavior tendency, but because it is difficult for the user to generate transaction behavior in a stable cycle, for example, use fast Fourier transform to quickly calculate time-domain sequences. None of the periodic methods can be well applied to the analysis of user transaction behavior, so it is difficult to accurately obtain data such as user loyalty to merchants or platforms.

SUMMARY OF THE INVENTION

In order to solve the above-mentioned problems in the prior art, one purpose of the present application is to propose a method and device for analyzing user behavior cycle, which can more accurately analyze user behavior and evaluate users, and can guide merchants to adjust production according to the evaluation results, so that Provide corresponding customized services or products to target users, so as to improve the efficiency of production services and reduce the waste of resources caused by blind production.

In order to achieve the above purpose, an embodiment of the present application proposes a method for analyzing a user behavior cycle, including: acquiring a time interval sequence of user behavior; searching for a periodic time interval sub-period where fluctuations of adjacent elements in the time interval sequence satisfy a predetermined condition. sequence; calculate the scores of all the subsequences of the periodic time interval; perform a graded evaluation on the user according to the scores.

In order to achieve the above purpose, an embodiment of the present application proposes a device for analyzing a user behavior cycle, including: an acquisition unit for acquiring a time interval sequence of user behaviors; a subsequence unit for finding adjacent neighbors in the time interval sequence. A periodic time interval subsequence whose fluctuations of elements satisfy a predetermined condition; a calculation unit for calculating the scores of all the periodic time interval subsequences; and an evaluation unit for grading the user according to the scores.

It can be seen from the technical solutions provided by the above embodiments of the present application that the periodic characteristics of user behaviors can be found through the methods of the above embodiments of the present application, and the periodic behaviors can be scored, quantifying the process of evaluating users, and using digital The method describes the characteristics of users, which can more accurately analyze user behavior and evaluate users. According to the evaluation results, merchants can be guided to adjust production to provide corresponding customized services or products to target users, so as to improve the efficiency of production services and reduce costs. Waste of resources caused by blind production.

Additional aspects and advantages of the present application will be set forth, in part, in the following description, and in part will be apparent from the following description, or learned by practice of the present application.

Of course, implementing any product or method of the present application does not necessarily require achieving all of the above-mentioned advantages at the same time.

Description of drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only These are some embodiments described in this application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without any creative effort.

FIG. 1 shows a flowchart of a method for analyzing a user behavior cycle according to an embodiment of the present application;

FIG. 2 is a schematic structural diagram of an apparatus for analyzing a user behavior cycle according to an embodiment of the present application;

FIG. 3 shows a specific flowchart of a method for analyzing a user behavior cycle according to an embodiment of the present application;

FIG. 4 is a schematic diagram of a time series of user consumption behaviors according to an embodiment of the present application;

FIG. 5 shows a specific flowchart of a method for analyzing a user behavior cycle according to an embodiment of the present application.

Detailed ways

Embodiments of the present application provide a method and apparatus for analyzing a user behavior cycle.

In order to make those skilled in the art better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described The embodiments are only a part of the embodiments of the present application, but not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative work shall fall within the scope of protection of this application.

FIG. 1 is a flowchart of a method for analyzing a user behavior cycle according to an embodiment of the present application. In this figure, it is described that the user behavior cycle is obtained by analyzing the user behavior, and the user is analyzed and evaluated according to the user behavior cycle. Methods. In the prior art, through the EFM model, where R represents the most recent consumption (Recency), F represents the consumption frequency (Frequency), and M represents the consumption amount (Monetary), these three indicators are mainly used to analyze the user's loyalty to the merchant or platform. , although this method can describe the user's behavior to a certain extent, it cannot accurately describe the user's behavior distribution on the entire time axis. Moreover, due to the discreteness of user behavior, the fast Fourier transform in the prior art cannot be well applied to the analysis of the user behavior cycle. Through the method in the embodiment of the present application, it is determined whether the fluctuation degree of the user behavior time interval constituting the sequence exceeds the threshold value, to find the subsequences in the sequence in periodic form, and to score all the subsequences, so as to obtain Analysis results for this user.

The method specifically includes step 101, obtaining a time interval sequence of user behaviors.

Step 102: Find a periodic time interval subsequence whose fluctuations of adjacent elements in the time interval sequence satisfy a predetermined condition.

Step 103: Calculate the scores of all the periodic time interval subsequences.

Step 104: Perform a graded evaluation on the user according to the score.

As an embodiment of the present application, the searching for a periodic time interval subsequence whose fluctuations of adjacent elements in the time interval sequence satisfy a predetermined condition further includes, when a certain continuous subsequence of the time interval sequence is adjacent to When the fluctuation of the element is less than the predetermined value, the continuous subsequence is a periodic time interval subsequence.

Wherein, the time interval sequence of user behaviors can be, for example, the user has consumption behaviors at KFC on January 3, January 7, and January 12, and with January 1 as the starting date, the corresponding user behaviors are generated. The time series is (2, 6, 11), that is, the time difference between the consumption date and the start date is calculated, and the time interval sequence of the user behavior is the interval between the user behavior times in the user behavior time series, that is, It is said that the time interval sequence of user behavior is (4, 5), and the element refers to each time interval in the time interval sequence.

As an embodiment of the present application, the fluctuation includes the variance of adjacent elements in the subsequence, or the ratio of adjacent elements in the subsequence.

Among them, for the ratio of adjacent elements, the larger of the two adjacent elements can be used as the numerator, and the smaller one can be used as the denominator, so that the ratio of the two adjacent elements can be obtained, and the judgment of the fluctuation of adjacent elements can also be used. For other methods in the prior art, the embodiments of the present application only enumerate several methods, and do not limit which method is used to obtain the fluctuation of adjacent elements.

As an embodiment of the present application, when the fluctuation of adjacent elements in a certain continuous subsequence of the time interval sequence is less than a predetermined value, the continuous subsequence is a periodic time interval subsequence further comprising: One element in the sequence is the head, and the other element is the tail to form a continuous subsequence. Calculate whether the current subsequence is a periodic time interval subsequence. If the subsequence is a periodic time interval subsequence, no longer calculate the subsequence Whether there are other periodic time interval subsequences.

Among them, calculating whether a subsequence is a periodic time interval subsequence is a traversal process. If a subsequence is a periodic time interval subsequence, any segment of the subsequence is a periodic time interval subsequence, so as to avoid searching for a periodic time interval subsequence. When a time interval subsequence is used, the subsequence with shorter periodic time interval is repeatedly searched and calculated in the subsequence, thereby reducing the computational complexity.

As an embodiment of the present application, forming a continuous subsequence with a certain element in the time interval sequence as the head and another element as the tail further includes, when selecting the head element and the tail element in the subsequence, taking the difference between the two elements. The more elements you have, the better.

Among them, the head and tail elements of the entire time interval sequence can be selected as the head and tail elements in the subsequence, and then the head and tail elements in the subsequence become other elements for traversal. By selecting and traversing a long sequence, the traversal can be reduced. The number of calculations can further reduce the computational complexity.

As an embodiment of the present application, when the fluctuation of adjacent elements in a certain continuous subsequence of the time interval sequence is less than a predetermined value, the continuous subsequence is a periodic time interval subsequence further comprising, the continuous subsequence If the subsequence satisfies the following formula, it is a periodic time interval subsequence:

Variance(list)/mean(list)<epsilon/f(length(list))

Among them, Variance(list) refers to the fluctuation of adjacent elements in the subsequence, mean(list) refers to the mean value of the elements in the subsequence, epsilon refers to the preset value according to the scene, and f(length(list)) refers to the consideration Penalty to subsequence length.

Among them, for example, the subsequence is (1, 1, 2, 2, 4), then the mean(list) is (1+1+2+2+4)/5=2, and the value of epsilon varies according to the scene , for example, epsilon may take 1 in the scenario of user consumption behavior, and epsilon may take 2 or other values in the scenario where the user pays water and electricity bills; Different operations are used in different application scenarios. length(list) refers to the length of the subsequence, and f(length(list)), for example, may be the root operation of the length of the subsequence, or other operations, or equal to the length of the subsequence.

As an embodiment of the present application, when the fluctuation of adjacent elements in a certain continuous subsequence of the time interval sequence is less than a predetermined value, the continuous subsequence is a periodic time interval subsequence further comprising, the continuous subsequence If the subsequence satisfies the following formula, it is a periodic time interval subsequence:

Variance(list)＜epsilon

Among them, Variance(list) refers to the fluctuation of adjacent elements in the subsequence, and epsilon refers to the preset value according to the scene.

As an embodiment of the present application, calculating the scores of all the periodic time interval subsequences further includes: extracting the period, fluctuation and length of the periodic time interval subsequences, assigning different weight values, and obtaining the period The time information from the end of the time interval subsequence to the current time is calculated to obtain the score of the periodic time interval subsequence, and the sum of the scores of all periodic time interval subsequences is calculated.

As an embodiment of the present application, calculating the score of the periodic time interval subsequence further includes that the score of the periodic time interval subsequence conforms to the following formula:

s=(a*period+b*variance+c*times)*f(toNow)

Among them, s is the score of the periodic time subsequence, a, b, and c are the weights, period is the period of the periodic time subsequence, variance is the fluctuation of adjacent elements in the periodic time subsequence, and times is the is the number of elements of the periodic time subsequence; toNow is the time information from the end of the periodic time interval subsequence to the current time, and f() is the function operation on the time information.

The smaller the period, the more frequent the user behavior, the larger the value of this item should be, a*period can be a negative correlation operation, for example, y=1/x; the smaller the variance, the smaller the fluctuation, the value of this item should be The larger the value is, the b*variance can be a negative correlation operation, for example, y=1/x; the larger the times, the longer the period, the larger the value of this item should be, and the c*times can be a positive correlation operation, such as y=x; When the toNow is larger, it means that the value of the periodic time interval subsequence is lower, and the score of the subsequence should be lower, and f() can be a root operation, a log operation, an e- ^x operation, etc., Where x is time information, and the f() can be different according to different application scenarios.

As an embodiment of the present application, calculating the scores of all the periodic time interval subsequences further includes recording parameters of all the periodic time interval subsequences, including user ID, periodic time granularity, and the periodic time interval subsequences period, fluctuation, number of elements, period start time, period end time.

The recorded content may also include the object of user behavior, such as the merchant ID of online shopping, the merchant ID of real consumption behavior, etc., and the parameters of other periodic time interval subsequences may also be recorded for later data analysis.

Through the method of the above embodiments of the present application, the periodic characteristics of user behaviors can be found, and the periodic behaviors can be scored, the process of evaluating users can be quantified, and the characteristics of users can be described in a digital way, which enables more accurate analysis. User behavior and evaluation of users, according to the evaluation results, can guide the business to adjust production to provide corresponding customized services or products to target users, thereby improving the efficiency of production services and reducing the waste of resources caused by blind production.

FIG. 2 is a schematic structural diagram of an apparatus for analyzing user behavior cycle according to an embodiment of the present application. Analysis and evaluation, in which each functional unit can be realized by programmable logic device, or realized by computer or computer cluster, and each functional unit can be integrated in one computer, or a certain functional unit or a certain functional unit can be realized by multiple computers. The functions of several functional units.

The apparatus specifically includes an acquisition unit 201 for acquiring a time interval sequence of user behaviors.

The subsequence unit 202 is configured to search for a periodic time interval subsequence whose fluctuations of adjacent elements in the time interval sequence satisfy a predetermined condition.

The calculation unit 203 is configured to calculate the scores of all the subsequences of the periodic time interval.

The evaluation unit 204 is configured to perform a graded evaluation on the user according to the score.

As an embodiment of the present application, the subsequence unit is specifically configured to: when the fluctuation of adjacent elements in a certain continuous subsequence of the time interval sequence is less than a predetermined value, the continuous subsequence is a periodic time spacer sequence.

As an embodiment of the present application, the fluctuation includes the variance of adjacent elements in the subsequence, or the ratio of adjacent elements in the subsequence.

As an embodiment of the present application, the subsequence unit is specifically used to calculate whether the current subsequence is a periodic time interval subsequence by taking an element in the time interval sequence as the head and another element as the tail to form a continuous subsequence, If the subsequence is a periodic time interval subsequence, it is not calculated whether there are other periodic time interval subsequences in the subsequence.

As an embodiment of the present application, the subsequence unit is specifically configured to, when selecting the head element and the tail element in the subsequence, the more elements there are between the two, the better.

As an embodiment of the present application, the subsequence unit is specifically used for, the continuous subsequence satisfies the following formula to be a periodic time interval subsequence:

Variance(list)/mean(list)<epsilon/f(length(list))

Among them, Variance(list) refers to the fluctuation of adjacent elements in the subsequence, mean(list) refers to the mean value of the elements in the subsequence, epsilon refers to the preset value according to the scene, and f(length(list)) refers to the consideration Penalty to subsequence length.

As an embodiment of the present application, the subsequence unit is specifically used for, the continuous subsequence satisfies the following formula to be a periodic time interval subsequence:

Variance(list)＜epsilon

Among them, Variance(list) refers to the fluctuation of adjacent elements in the subsequence, and epsilon refers to the preset value according to the scene.

As an embodiment of the present application, the computing unit is specifically configured to extract the period, fluctuation and length of the periodic time interval subsequence, assign different weight values, and obtain the period from the end of the periodic time interval subsequence to the current The time information of the time, the score of the subsequence of the periodic time interval is calculated, and the sum of the scores of all the subsequences of the periodic time interval is calculated.

As an embodiment of the present application, the calculation unit is specifically configured to calculate and obtain the score of the periodic time interval subsequence further comprising, the score of the periodic time interval subsequence conforms to the following formula:

s=(a*period+b*variance+c*times)*f(toNow)

Among them, s is the score of the periodic time subsequence, a, b, and c are the weights, period is the period of the periodic time subsequence, variance is the fluctuation of adjacent elements in the periodic time subsequence, and times is the is the number of elements of the periodic time subsequence; toNow is the time information from the end of the periodic time interval subsequence to the current time, and f() is the function operation on the time information.

As an embodiment of the present application, the computing unit is further configured to record all the parameters of the periodic time interval subsequence, including user ID, periodic time granularity, period, fluctuation, element number of the periodic time interval subsequence number, cycle start time, cycle end time.

Through the device of the above embodiment of the present application, the periodic characteristics of user behavior can be found, and the periodic behavior can be scored, the process of user evaluation is quantified, the characteristics of users are described in a digital way, and more accurate analysis can be achieved User behavior and evaluation of users, according to the evaluation results, can guide the business to adjust production to provide corresponding customized services or products to target users, thereby improving the efficiency of production services and reducing the waste of resources caused by blind production.

Figure 3 shows a specific flowchart of a method for analyzing a user behavior cycle according to an embodiment of the present application. For a certain period of time, consumption is maintained once every other day, and then there is a discontinuity, and then consumption is once every two days. This consumption behavior cannot be found using the fast Fourier transform in the prior art. In this application In the solution of , the inventor provides a quasi-traversal period finding method, which can find the period subsequences in the user's consumption behavior, so that the user's behavior can be subdivided and users with corresponding behaviors can be graded and evaluated.

Step 301, acquiring a time series of user consumption behaviors.

In this step, as shown in Figure 4, the user's consumption behavior period is 1 at the beginning, and after 7 days, the user's consumption behavior period is 2, and the time series is (1, 2, 3 , 4, 6, 11, 13, 15), representing the user's consumption behavior occurred on the 1st, 2nd, 3rd, 4th, 6th, 11th, 13th, and 15th, respectively.

Step 302: Obtain a time interval sequence of user consumption behaviors according to the time series of user consumption behaviors.

Since the user's behavior is consumption behavior, the granularity of the time interval sequence is "days". According to the time series diagram of consumption behavior shown in Fig. 4, a time interval sequence (1, 1, 1, 2, 5, 2, 2) can be obtained, and the time interval sequence represents the time interval of the above-mentioned time series.

Step 303 , traverse all the subsequences in the time interval sequence, and find the periodic time interval subsequence therein.

In this step, all subsequences of the time interval sequence are traversed, and the traversal method must satisfy a principle: traverse the long subsequence first. If the long subsequence is a periodic sequence, then all the subsequences in the subsequence will be At the same time, if the current subsequence is a subsequence of any sequence in the periodic sequence set E, it is no longer judged whether it is a periodic sequence, which can effectively reduce the computational complexity, so this algorithm is named "quasi-periodic sequence". Traversal".

The specific process can refer to the following pseudo program:

Input: time series M

Output: Periodic time series set E, and corresponding period parameters

Among them, length(N) is the length of the time interval sequence, list is the subsequence, and E.append(list) is to put the list subsequence into the sequence set E. In this embodiment, through the loop, from the longest time The interval sequence starts to be judged (see the 5th line of the above program), and the elements in the time interval sequence are gradually reduced (reduced from the last element forward) to form a time interval subsequence (see the 4th line of the above program), and each time interval is judged. Whether the time interval subsequence is a periodic time interval subsequence, then move the head element in the time interval sequence backward (see the third line of the above program), and re-determine whether the current time interval subsequence is a periodic time interval subsequence, if If a certain time interval subsequence is a periodic time interval subsequence, it will not continue to judge whether the subsequence also includes other periodic time interval subsequences (see the seventh line of the above program).

The specific method for judging whether the subsequence is a periodic time interval subsequence may be: judging whether the variance of the time interval subsequence is less than a predetermined value, that is, Variance(list)<epsilon, where Variance(list) refers to the variance of the subsequence, epsilon It means that according to the preset value of the scene, it can be preset to 1 in the scene of the user's consumption behavior.

For example, in the first cycle, the time interval sequence is (1, 1, 1, 2, 5, 2, 2), and the mean of the time interval sequence is (1+1+1+2+5+2+2)/ 7=2, the variance of the time interval series is ((2-1) ² +(2-1) ² +(2-1) ² +(2-2) ² +(5-2) ² +(2- 2) ² +(2-2) ² )/7=12/7, the variance is greater than the preset value 1, and the time interval sequence is not a periodic time interval sequence; in a subsequent cycle, the time interval subsequence is ( 1, 1, 1, 2), the mean of the time interval subsequence is 5/4, the variance of the time interval subsequence is 3/16, and the variance is less than the preset value 1, the time interval subsequence (1, 1, 1, 2) are periodic time interval subsequences.

Another periodic time interval subsequence (2,2) can be found by a similar method.

In another embodiment, it is also possible to judge whether the time interval subsequence is a periodic time interval subsequence according to the following formula:

Variance(list)/mean(list)<epsilon/f(length(list))

Among them, Variance(list) refers to the variance of the subsequence, mean(list) refers to the mean value of the elements in the subsequence, epsilon refers to the preset value according to the scene, and f(length(list)) refers to considering the length of the subsequence penalty value.

Step 304: Calculate the scores of the periodic time interval subsequences found above.

In this step, the score of the periodic time interval subsequence can be calculated according to the weights of the parameters of the subsequence, for example, the score of the subsequence is calculated with reference to the period, fluctuation and length of the periodic time interval subsequence,

s=a*period+b*variance+c*times

Where, a, b, and c are weights, period is the period of the periodic time subsequence, variance is the fluctuation of the periodic time subsequence, and times is the number of elements of the periodic time subsequence. Among them, the smaller the period, the more frequent the user behavior, the larger the value of this item should be, a*period can be a negative correlation operation, such as y=1/x; the smaller the variance, the smaller the fluctuation, the value of this item should be The larger the value, b*variance can be a negative correlation operation, for example, y=1/x; the larger the times, the longer the period, the larger the value of this item should be, and c*times can be a positive correlation operation, for example, y=x.

In the same way, the scores of other periodic time interval subsequences are calculated, and the scores of all periodic time interval subsequences are added to obtain the score of the time interval sequence.

In another embodiment, the penalty value from the end of the periodic time interval subsequence to the current time may also be considered, that is, the longer the end time of the periodic time interval subsequence to the current time, the score should be relatively low. The shorter the interval subsequence end time to the current time, the higher the score should be:

s=(a*period+b*variance+c*times)*f(toNow)

Wherein, toNow is the time information from the end of the periodic time interval subsequence to the current time, and f( ) is the function operation on the time information, which can be root operation, log operation, e- ^x operation, etc., wherein x is time information, and the f() can be different according to different application scenarios.

The time interval subsequences of the user's consumption behavior can be stored separately, for example, can be stored in Table 1 shown in the following table, and evaluation by other user evaluation algorithms can be facilitated according to the table.

Table 1

Step 305: Perform a graded evaluation on the user according to the score.

According to the parameters of the sub-sequences of each cycle time interval, the user's loyalty can be obtained by analysis, for example, the higher the score, the higher the user loyalty, and the user groups can be clustered according to the various parameters in the above table, for example Identical or similar users are divided into one category, and graded according to the fluctuation of user consumption behavior. After users are classified or graded, different services can be provided to users according to the classification or grade, which can include sending different information to users, giving customers Different discounts, etc., that is to say, users can be graded and evaluated through the foregoing steps, and then various services can be provided to users by using the above data.

FIG. 5 is a specific flowchart of a method for analyzing a user behavior cycle according to an embodiment of the present application. In this figure, a method for analyzing a user is described by taking the behavior of a user paying electricity bills as an example. In this embodiment, for the sake of example , it is assumed that the user may keep paying the electricity bill every other week for a certain period of time, and then there will be discontinuities, and then the bill will be paid every two weeks. This kind of payment behavior cannot be found using the fast Fourier transform in the prior art. Regularly, in the solution of the present application, the inventor provides a quasi-traversal period search method, which can find the period subsequence in the user's electricity bill payment behavior, so that the user's behavior can be subdivided and users with corresponding behavior can be classified. Evaluation.

Step 501: Obtain a time series of user payment behaviors.

In this step, in order to explain the time series simply, we still take Figure 4 as an example. , Week 6, Week 11, Week 13, Week 15.

Step 502: Obtain a time interval sequence of user payment behaviors according to the time sequence of user payment behaviors.

In this step, since the user's behavior is payment behavior, the granularity of the time interval sequence is "week". A time interval sequence (1, 1, 1, 2, 5, 2, 2) can be obtained, and the time interval sequence represents the time interval of the above-mentioned time series.

Step 503, traverse all the subsequences in the time interval sequence, and find the periodic time interval subsequence therein.

In this step, the "greedy method" can be used to traverse the time interval sequence to find the periodic time interval subsequence, or directly traverse the time interval sequence to find the periodic time interval subsequence.

The "greedy method" means that when solving a problem, always make the best choice at the moment. The "greedy method" is used to extract the subsequences of the sequence, which effectively reduces the computational time complexity, but will miss some periodic subsequences.

For the sake of simplicity, in this embodiment, the method of direct traversal is used to find the periodic time interval subsequence, and the fluctuation between adjacent elements is judged from the head element of the time interval sequence to the next element by one element. If it exceeds a predetermined value, Then take the traversed time interval subsequence as the periodic time interval subsequence, and start from the element whose fluctuation exceeds the predetermined value, and continue to search backwards for the subsequence whose adjacent element fluctuation is less than the predetermined value.

In this embodiment, the fluctuation between adjacent elements can be judged by the ratio between the two elements. When the ratio of the two elements is large, it is considered that the fluctuation of the two elements is large, which may be periodic When the ratio is small, it is considered that the fluctuation between the two elements is small, and there may be periodicity.

According to the above fluctuation, whether the last element and the previously traversed element subsequence constitute a periodic time interval subsequence can be obtained directly according to the comparison result between the ratio of the adjacent elements and the predetermined value.

Step 504: Calculate the scores of the periodic time interval subsequences found above.

For this step, reference may be made to the corresponding steps in the above-mentioned embodiment shown in FIG. 3 , which will not be repeated here.

Step 505: Perform a graded evaluation on the user according to the score.

Through the method and device in the above embodiments of the present application, the periodic characteristics of user behavior can be found, and the periodic behavior can be scored, the process of evaluating users can be quantified, the characteristics of users can be described in a digital way, and the user can be more Accurately analyze user behavior and evaluate users. According to the evaluation results, merchants can be guided to adjust production to provide corresponding customized services or products to target users, thereby improving the efficiency of production services and reducing the waste of resources caused by blind production.

The improvement of a technology can be clearly distinguished from improvement in hardware (for example, improvement of circuit structures such as diodes, transistors, switches, etc.) or improvement in software (improvement in method flow). However, with the development of technology, the improvement of many methods and processes today can be regarded as a direct improvement of the hardware circuit structure. Designers almost get the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that the improvement of a method flow cannot be realized by hardware entity modules. For example, a Programmable Logic Device (PLD) (eg, Field Programmable Gate Array (FPGA)) is an integrated circuit whose logic function is determined by user programming of the device. It is programmed by the designer to "integrate" a digital system on a PLD without having to ask a chip manufacturer to design and manufacture a dedicated integrated circuit chip 2 . And, instead of making integrated circuit chips by hand, these days, most of this programming is done using "logic compiler" software, which is similar to the software compilers used in program development and writing, but before compiling The original code also has to be written in a specific programming language, which is called Hardware Description Language (HDL), and there is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (JavaHardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. Currently, the most commonly used is VHDL ( Very-High-Speed IntegratedCircuit Hardware Description Language) and Verilog2. It should also be clear to those skilled in the art that a hardware circuit for implementing the logic method process can be easily obtained by simply programming the method process in the above-mentioned several hardware description languages and programming it into the integrated circuit.

The controller may be implemented in any suitable manner, for example, the controller may take the form of eg a microprocessor or processor and a computer readable medium storing computer readable program code (eg software or firmware) executable by the (micro)processor , logic gates, switches, application specific integrated circuits (ASICs), programmable logic controllers and embedded microcontrollers, examples of controllers include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20 and Silicon Labs C8051F320, the memory controller can also be implemented as part of the control logic of the memory.

Those skilled in the art also know that, in addition to implementing the controller in the form of pure computer-readable program code, the controller can be implemented as logic gates, switches, application-specific integrated circuits, programmable logic controllers and embedded devices by logically programming the method steps. The same function can be realized in the form of a microcontroller, etc. Therefore, such a controller can be regarded as a hardware component, and the devices included therein for realizing various functions can also be regarded as a structure within the hardware component. Or even, the means for implementing various functions can be regarded as both a software module implementing a method and a structure within a hardware component.

The systems, devices, modules or units described in the above embodiments may be specifically implemented by computer chips or entities, or by products with certain functions.

For the convenience of description, when describing the above device, the functions are divided into various units and described respectively. Of course, when implementing the present application, the functions of each unit may be implemented in one or more software and/or hardware.

From the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on this understanding, the technical solutions of the present application can be embodied in the form of software products in essence or the parts that make contributions to the prior art, and the computer software products can be stored in storage media, such as ROM/RAM, magnetic disks , CD-ROM, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present application.

Each embodiment in this specification is described in a progressive manner, and the same and similar parts between the various embodiments may be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, as for the system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and for related parts, please refer to the partial descriptions of the method embodiments.

The present application may be used in numerous general purpose or special purpose computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, including A distributed computing environment for any of the above systems or devices, and the like.

The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including storage devices.

Although the application has been described by way of examples, those of ordinary skill in the art will recognize that the application is subject to many modifications and variations without departing from the spirit of the application, and the appended claims are intended to include such modifications and changes without departing from the spirit of the application.

Claims (20)

1. A method of analyzing a period of user behavior, comprising:

acquiring a time interval sequence of user behaviors;

searching for a periodic time interval subsequence in which the fluctuation of adjacent elements in the time interval sequence meets a predetermined condition;

calculating scores for all of the periodic time interval subsequences;

grading evaluation is carried out on the user according to the score;

calculating scores for all of the periodic time interval subsequences further comprises: extracting the period, fluctuation and length of the periodic time interval subsequence, distributing different weight values, obtaining time information from the end of the periodic time interval subsequence to the current time, and calculating to obtain the score of the periodic time interval subsequence.

2. The method of claim 1, wherein finding a periodic time interval subsequence in which fluctuations of neighboring elements in the time interval sequence satisfy a predetermined condition further comprises, when fluctuations of neighboring elements in a certain continuous subsequence of the time interval sequence are less than a predetermined value, then the continuous subsequence is a periodic time interval subsequence.

3. The method of claim 2, wherein the fluctuation comprises a variance of adjacent elements in the subsequence, or a ratio of adjacent elements in the subsequence.

4. The method according to claim 2, wherein when the fluctuation of adjacent elements in a certain continuous subsequence of the time interval sequence is smaller than a predetermined value, the continuous subsequence is a periodic time interval subsequence, and the step of calculating whether the current subsequence is a periodic time interval subsequence is further included.

5. The method of claim 4, wherein forming a continuous subsequence with an element in the sequence of time intervals as a head and another element as a tail further comprises selecting the subsequence with the head element and the tail element as having the more elements in between.

6. The method of claim 2, wherein when the fluctuation of adjacent elements in a continuous subsequence of the time interval sequence is less than a predetermined value, then the continuous subsequence is a periodic time interval subsequence further comprising, if the continuous subsequence satisfies the following formula:

Variance(list)/mean(list)＜epsilon/f(length(list))

where, variance (list) refers to fluctuation of adjacent elements in the subsequence, mean (list) refers to a mean value of elements in the subsequence, epsilon refers to a preset value according to a scene, and f (length (list)) refers to a penalty value considering the length of the subsequence.

7. The method of claim 2, wherein when the fluctuation of adjacent elements in a continuous subsequence of the time interval sequence is less than a predetermined value, then the continuous subsequence is a periodic time interval subsequence further comprising, if the continuous subsequence satisfies the following formula:

Variance(list)＜epsilon

wherein, variance (list) refers to the fluctuation of adjacent elements in the subsequence, and epsilon refers to a preset value according to a scene.

8. The method of claim 1, wherein calculating scores for all of the periodic time interval subsequences further comprises calculating a sum of scores for all of the periodic time interval subsequences.

9. The method of claim 8, wherein calculating the score for the periodic time interval subsequence further comprises calculating the score for the periodic time interval subsequence according to the formula:

s＝(a*period+b*variance+c*times)*f(toNow)

wherein s is the score of the periodic time subsequence, a, b and c are weights, period is the period of the periodic time subsequence, variance is the fluctuation of adjacent elements in the periodic time subsequence, and times is the number of the elements of the periodic time subsequence; toNow is time information from the end of the periodic time interval subsequence to the current time, and f () is a function operation on the time information.

10. The method of claim 1, wherein calculating scores for all of the periodic time interval subsequences further comprises recording parameters of all of the periodic time interval subsequences including user ID, periodic time granularity, period, fluctuation, element count, period start time, and period end time of the periodic time interval subsequences.

11. An apparatus for analyzing a period of user behavior, comprising:

the acquiring unit is used for acquiring a time interval sequence of user behaviors;

a sub-sequence unit for finding a periodic time interval sub-sequence in which fluctuations of adjacent elements in the time interval sequence satisfy a predetermined condition;

a calculating unit, configured to calculate scores of all the periodic time interval subsequences;

the evaluation unit is used for carrying out grading evaluation on the user according to the score;

the calculating unit is specifically configured to extract a period, fluctuation, and length of the periodic time interval subsequence, assign different weight values, obtain time information from the end of the periodic time interval subsequence to the current time, and calculate a score of the periodic time interval subsequence.

12. The apparatus according to claim 11, wherein the subsequence unit is configured to determine a period time interval subsequence if the fluctuation of adjacent elements in a certain continuous subsequence of the time interval sequence is smaller than a predetermined value.

13. The apparatus of claim 12, wherein the fluctuation comprises a variance of adjacent elements in the subsequence or a ratio of adjacent elements in the subsequence.

14. The apparatus according to claim 12, wherein the sub-sequence unit is specifically configured to form a continuous sub-sequence with an element in the time interval sequence as a head and another element as a tail, calculate whether the current sub-sequence is a periodic time interval sub-sequence, and if the sub-sequence is the periodic time interval sub-sequence, not calculate whether there are other periodic time interval sub-sequences in the sub-sequence.

15. The apparatus according to claim 14, wherein the subsequence unit is specifically configured to select the first element and the last element of the subsequence as having more elements in between.

16. The apparatus according to claim 12, wherein the subsequence unit is specifically configured to determine the consecutive subsequences as periodic time interval subsequences if the following formula is satisfied:

Variance(list)/mean(list)＜epsilon/f(length(list))

where, variance (list) refers to fluctuation of adjacent elements in the subsequence, mean (list) refers to a mean value of elements in the subsequence, epsilon refers to a preset value according to a scene, and f (length (list)) refers to a penalty value considering the length of the subsequence.

17. The apparatus according to claim 12, wherein the subsequence unit is specifically configured to determine the consecutive subsequences as periodic time interval subsequences if the following formula is satisfied:

Variance(list)＜epsilon

wherein, variance (list) refers to the fluctuation of adjacent elements in the subsequence, and epsilon refers to a preset value according to a scene.

18. The apparatus according to claim 11, wherein the calculation unit is specifically configured to calculate a sum of all periodic time interval subsequence scores.

19. The apparatus according to claim 18, wherein the calculating unit is specifically configured to calculate the score of the periodic time interval subsequence further includes that the score of the periodic time interval subsequence satisfies the following formula:

s＝(a*period+b*variance+c*times)*f(toNow)

wherein s is the score of the periodic time subsequence, a, b and c are weights, period is the period of the periodic time subsequence, variance is the fluctuation of adjacent elements in the periodic time subsequence, and times is the number of the elements of the periodic time subsequence; toNow is time information from the end of the periodic time interval subsequence to the current time, and f () is a function operation on the time information.

20. The apparatus according to claim 11, wherein the computing unit is further configured to record parameters of all the periodic time interval subsequences, including user ID, periodic time granularity, period, fluctuation, element number, period start time, and period end time of the periodic time interval subsequences.

Images