JP2014115908A - Word search device and word search method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a word search device and a word search method that are capable of determining unknown words from words included in a text, excluding known words and noise words, and showing the meaning of a word that a user wants to know.SOLUTION: The word search device shows only a word of which a user wants to know by excluding known words and noise words from words in sounds and a document and showing unknown words alone. Further, the word search device shows words that are needed to be showed highly by including words that first appear and words that have already appeared and are frequently referred to by a user into the unknown words.

Description

  The present invention relates to a word search apparatus and a word search method for detecting a word from text such as speech or characters and displaying the meaning of the detected word.

  Conventionally, when there is an unknown word in text such as speech or letters, it has been necessary to look up the unknown word in a dictionary or search on an Internet search site. As one means for saving this search effort, there is a means for displaying the meaning of a word included in a sentence including a designated character in a separate column when a desired character in the text is designated. In this method, a sentence including a selected character is subjected to morphological analysis, divided into units of words, each word is searched as a search keyword, and the result is displayed (for example, see Patent Document 1).

JP 2011-181109 A

  In the conventional technology described above, all the words included in the specified sentence are searched, so only the word that the user wants to know is searched (without searching for unnecessary words) and the meaning is displayed. It was difficult to do.

  The present invention has been made in view of the above conventional example, and an object thereof is to provide a word search device and a word search method that can display the meaning of a word desired by a user, excluding known words and noise.

In order to achieve the above object, the present invention has the following configuration. That is,
A word search device that determines an unknown word from words included in a text and outputs a word list of unknown words,
Word storage means for storing noise words and known words;
Search means for searching a dictionary for words that do not correspond to the noise word or the known word among the words included in the text;
A word registration unit for registering a word for which the search by the search unit has failed as the noise word, and newly registering a word for which the search has been successful as the known word;
Creating means for creating a word list including, as unknown words, words newly registered as the known words from the words included in the text;
Means for transmitting the word list to a terminal for display.

  According to the present invention, only words whose meanings are to be known are searched for except for known words and noise words, and the meanings are displayed. Therefore, it is possible to display the words necessary for the user and their meanings.

It is a figure which shows the whole block diagram of embodiment suitable for this invention. 2 is a diagram illustrating an example of a hardware configuration of a portable device 102. FIG. It is a figure which shows an example of the hardware constitutions of the document analysis apparatus 103 and the word dictionary presentation apparatus 104. 1 is a system configuration diagram of a portable device 102. FIG. 2 is a system configuration diagram of a document analysis apparatus 103. FIG. It is an example of the word which the word preservation | save part 504 preserve | saves. It is an example of the result which the access analysis preservation | save part 508 preserve | saves. 6 is an example of a display result output by the portable device. 2 is a system configuration diagram of a word dictionary presentation device 104. FIG. It is the whole flow for implementing this embodiment. It is a detailed unknown word judgment flow of processing S1003 for judging an unknown word. It is a detailed flow of a semantic search and word registration process S1004. It is a detailed flow of process S1005 which displays the meaning of a word. It is a figure which shows the procedure which performs determination of a well-known word. It is a figure which shows an example of UI for designating a level.

  The best mode for carrying out the present invention will be described below with reference to the drawings.

[Embodiment 1]
<System hardware configuration>
FIG. 1 is a diagram showing an overall configuration diagram of a document analysis system or a word search system according to an embodiment suitable for the present invention. FIG. 1 shows a state in which a portable device 102 functioning as a terminal is connected to each server via the Internet 101. Assume that each of the portable devices 102 is owned by a plurality of users, and a plurality of portable devices 102 exist in the system of FIG. Each server is a document analysis device 103 and a word dictionary presentation device 104.

  The Internet 101 is a communication line for exchanging information between the above-described devices across a firewall. Via the Internet 101, communication from the portable device 102 to the servers 103 and 104 is possible through the firewall. The Internet 101 is a communication network that supports, for example, the TCP / IP protocol and can be wired or wireless. In FIG. 1 of the present embodiment, the servers 103 and 104 are shown as two servers. However, the servers 103 and 104 may be composed of one server computer, or may be composed of three or more server computers. It doesn't matter.

  FIG. 2 is a diagram illustrating an example of a hardware configuration of the portable device 102. The CPU 201 is a central processing unit that is a control unit of the portable device 102, executes an application program and OS stored in the HDD 205, and temporarily stores information, files, and the like necessary for executing the program in the RAM 204. Take control. The ROM 202 is a read-only memory that is a storage unit, and stores part or all of an operation system program that is a control program and various data.

  The NIC 203 is a network interface, and executes communication control processing with other devices via the interface. The RAM 204 is an occasional access memory that is temporary storage means, and functions as a main memory and work area of the CPU 201. The HDD 205 is one of external storage means and is a hard disk (HD) used as a large-capacity memory, and stores an application program, an OS, and related programs.

  The display 206 is an output device that is a display means and an input device that is an input means. A touch sensor is mounted on the display 206, and a program displayed by touch input can be operated. In this embodiment, the input device is described as being capable of inputting by touching the display. However, an input device may be provided by connecting a keyboard or a numeric keypad separately. Further, by touching the displayed text, it is possible to specify a character or a sentence including the character.

  A system bus 207 is a bus that connects components in the portable device 102 and controls the flow of data in the portable device 102. The microphone unit 208 collects ambient sounds, sounds, and the like as digital signals. The digital voice signal is recorded on a hard disk or the like, and can be converted into text by voice recognition processing.

  FIG. 3 is a diagram illustrating an example of the hardware configuration of the document analysis device 103 and the word dictionary presentation device 104. The document analysis device 103 and the word dictionary presentation device 104 include a CPU 301, ROM 302, NIC 303, keyboard 304, RAM 305, HDD 306, display 307, and system bus 308. The keyboard 304 is an input device that is an instruction input unit, and inputs an instruction of a control command for an application in the document analysis device 103 and the word dictionary presentation device 104. The display 307 is an output device that is a display unit, and displays commands input from the keyboard 304, the status of programs, and the like. Since the role of each other device is the same as that of the portable device 102, the detailed description is omitted.

<Configuration by function of each device>
4 is a system configuration diagram of the portable device 102 shown in FIG. 1, and functions as the processing units shown in FIG. 4 through cooperation of the hardware shown in FIG.
The voice input reception unit 401 is a part that receives a signal collected by the microphone unit 208 as a digital signal. The voice character string conversion unit 402 performs voice recognition processing on the signal received by the voice input reception unit 401 to convert it into text. Regarding the technology for converting speech signals into text (speech recognition processing), since a known technology can be used, a detailed description is omitted. The communication unit 403 sends the text created by the speech string conversion unit 402 to the document analysis apparatus 103 together with the analysis request, and receives the analysis result. The display unit 404 displays the analysis result received from the communication unit 403 on the display 206. The operation unit 405 can perform an operation on the program displayed on the display 206, and transmits the operation result to the document analysis apparatus 103 through the communication unit 403.

  FIG. 5 is a system configuration diagram of the document analysis apparatus 103 shown in FIG. The communication unit 501 receives the data transmitted from the portable device 102 by the NIC 303, transmits the semantic search from the semantic search unit 505 to the word dictionary presenting device 104, and the result generated by the result output unit 509 is portable. To the device 102.

  The morphological analysis unit 502 performs morphological analysis on the input text and extracts words included in the text. Since a known technique can be applied to a specific morphological analysis processing algorithm, it will not be described in detail here.

  The unknown word determination unit 503 determines whether the word extracted by the morphological analysis unit 502 is an unknown word, a known word, or a noise word. An unknown word refers to a word that has not been searched for meaning as a word and is not registered in the word storage unit 504. That is, it is a word that appears for the first time and can be rephrased as the first word. The known word has a meaning as a word, and indicates a word registered in the word storage unit 504 together with the meaning as a known word. In other words, it can be rephrased as an already-presented word. A noise word is a word registered in the word storage unit 504 as a noise word. Noise words are words that do not need to know their meanings, and include, for example, particles and conjunctions. An unknown word is searched for as a word by the word dictionary presentation device 104. As a result, a word that has not been registered in the dictionary is determined to be a noise word that has no meaning, and is registered in the word storage unit 504. As a result of searching for an unknown word, if it is registered in the dictionary, it is registered as a known word in the word storage unit 504 in association with its meaning.

  The word storage unit 504 is a storage area for storing known words and noise words, and is a word storage unit. An example of words stored by the word storage unit 504 is shown in FIG. FIG. 6A is an example of words registered as known words. The known word is registered in association with the word text, its meaning, and the word ID. FIG. 6B is an example of a word registered as a noise word. Since it is only necessary to determine that a noise word is a noise word, it is sufficient that text is included.

  The meaning search unit 505 is a part that executes word meaning search. The meaning retrieval unit 505 inquires of the word dictionary presenting device 104 whether or not the designated word has a meaning and, if it has a meaning, the meaning. The inquiry result is registered in the DB registration unit 506 as a known word or a noise word depending on whether there is a meaning returned from the word dictionary presentation device 104. In this embodiment, the semantic search is executed by another server and the search result is registered. However, the semantic search unit has a dictionary in which the meaning of the word is registered, and the search result is stored in the DB registration unit. In addition to the word dictionary presentation device 104, the meaning of the word may be inquired to a plurality of servers.

  The access analysis unit 507 stores, in the access analysis storage unit 508, the number of appearances and the number of references of words registered as known words among the words transmitted from the portable device 102. Thereafter, the access analysis unit 507 outputs the word to be output and the meaning to the portable device 102 from the analysis result to the result output unit 509.

  The access analysis storage unit 508 is an area for storing results retrieved from the portable device 102. An example of the stored result is shown in FIG.

  In the list of FIG. 7, tenant ID, user ID, word ID, appearance count, and reference count are stored. The tenant ID is an ID indicating a user group that bundles a plurality of users placed under the same management form in a multi-tenant environment. The user ID is an ID that uniquely indicates a user in one tenant. Therefore, the user can be uniquely indicated by the combination of the tenant ID and the user ID. The word ID is an ID indicating the word and meaning shown in FIG. The number of appearances is a value that counts how many times the word managed by the word ID has appeared in the past for each user. The reference count is a value obtained by counting the number of times the user has referred to each user's portable device 102 with respect to the appearing word. FIG. 7 is a table in which the number of appearances of words and the number of references for each user is recorded, and is referred to as a reference rate table.

  Multi-tenant refers to an environment where equipment and databases are shared by multiple customers. By using a multi-tenant, the same service can be provided to different customers as different services. The tenant ID in the present embodiment may be regarded as a group that bundles users.

  The result output unit 509 includes words other than noise words from the input words (known words registered in the word storage unit 504 and unknown words newly searched by the semantic search unit 505), and the meanings of those words. Is formatted into a format suitable for display on the portable device 102 and output. An example of a suitable format is shown in FIG. FIG. 8 is a diagram illustrating an example of a state displayed on the portable device 102. This example is an example of a UI that displays a list of words other than noise words from the input words and displays the meaning of the selected word when a user-desired word is selected from the list. The display is an example, and the display is not limited as long as the word and its meaning are displayed. The result output unit 509 transmits the output result formed through the communication unit 501 to the portable device 102.

  As described above, the document analysis apparatus 103 has a function of performing a word search and presenting it together with a meaning, and can also be called a word search apparatus.

  FIG. 9 is a system configuration diagram of the word dictionary presentation device 104 shown in FIG. The communication unit 901 transmits the word received from the document analysis device 103 to the word confirmation unit 902, and transmits the search result using the dictionary to the document analysis device 103. The word confirmation unit 902 inquires the dictionary unit 903 about the meaning of the word received from the communication unit. If the word is registered in the dictionary unit 903, the word confirmation unit 902 responds the word and the meaning. If not registered, the word confirmation unit 902 registers the word. The fact that there is no message is transmitted to the document analysis apparatus 103 via the communication unit 901. The dictionary unit 903 is a database that stores words and meanings in association with each other.

<Procedure for displaying meaning of words in text>
Next, means for outputting words and their meanings in this embodiment will be described with reference to the flowcharts of FIGS.

  FIG. 10 is an overall flow for carrying out this embodiment. The portable device 102 receives the voice signal at the microphone unit 208 and transmits the result of the voice recognition process as text to the document analysis apparatus 103 (S1001). Upon receiving the text, the document analysis apparatus 103 extracts words from the text by the morphological analysis unit 502 (S1002). The extracted word is determined to be an unknown word or a known word (S1003). If it is an unknown word, its meaning is searched, and a word having a meaning and a noise word are stored in the word storage unit 504 (S1004). The portable device 102 displays the result of searching for the meaning of the meaningful word (S1005). In this example, when a plurality of identical words are included in the text, the plurality of identical words are treated as one word. However, in order to record in the reference table etc. of FIG. 7 mentioned later, you may memorize | store the frequency | count of appearance for every word which appeared in the text of processing object linked | related with a word. In this case, in S1306 described later, the number of appearances in the stored text is added to the number of appearances instead of 1.

<Unknown word judgment procedure>
FIG. 11 is a detailed unknown word determination flow of process S1003 for determining an unknown word. The document analysis apparatus 103 compares the extracted word with each registered noise word (S1101), and determines whether the extracted word includes a word included in the noise word (S1102). If there is a word included in the noise word, the word is removed from the extracted word group (S1103). The remaining words are compared with each registered known word (S1104), and it is determined whether or not it is a registered known word (S1105). The words determined to be known words are stored in the access analysis storage unit 508 as appearing words that appear in the text to be analyzed (S1106), and the process proceeds to the display processing in S1005. On the other hand, for a word determined not to be a known word, the process proceeds to a semantic search and word registration process in S1004 (details will be described later with reference to FIG. 12).

<Semantic search and word registration procedure>
FIG. 12 is a detailed flow of the semantic search and word registration processing S1004. The meaning search unit 505 of the document analysis device 103 acquires a reference destination definition for searching for meaning (S1201), and inquires the word dictionary presenting device 104 about the meaning of the word (S1202). The reference destination definition for searching the meaning is, for example, an address of the word dictionary presentation device 104, and the inquiry is configured by combining a word or a group of words with a message indicating that the meaning of the word is inquired, for example. The word dictionary presenting apparatus 104 that has received the inquiry searches the meaning of the inquired word in a database in which the word and its meaning are registered (S1203). If the meaning cannot be searched, that is, the word does not hit the database. If the search is not successful, the fact that the search has failed is returned to the semantic search unit 505 of the document analysis apparatus 103. If the search fails, the semantic search unit 505 of the document analysis apparatus 103 determines whether or not there is a next search destination, and if so, inquires the next search destination about the meaning of the same word again (S1206). . When there is no next search destination (that is, when all search destinations have been searched), the inquired word is stored as a noise word from the DB registration unit 506 to the word storage unit 504 (S1205). When the meaning of the word can be searched (that is, when the search is successful), the meaning of the word is returned to the meaning search unit 505, and the meaning search unit 505 determines the searched word and its meaning as a new known word. Registration is performed in the word storage unit 504 via the DB registration unit 506 (S1207). Then, the word registered as a known word is temporarily stored as an appearance word appearing in the text to be analyzed in addition to the appearance word registered in step S1106 (S1208), and the process proceeds to the display processing of S1005.

<Display procedure>
FIG. 13 is a detailed flow of processing S1005 for displaying the meaning of a word. The access analysis unit 507 of the document analysis apparatus 103 acquires the appearance word temporarily stored in S1106 or S1208 (S1301), acquires the reference rate table (FIG. 7) from the access analysis storage unit 508 (S1302), and refers to it. An attempt is made to register an appearance word in the rate table (S1303). When a request for registration of the appearance word is made, it is determined whether or not each word has already been registered (S1304). If it is an unregistered word, the ID, the number of appearances, and the tenant ID of the appearance word are displayed in the reference rate table. The user ID is registered (S1305). Since no reference is made at the time of new registration, 0 is set as the reference count. As the tenant ID and user ID, for example, the ID included in the analysis request received together with the text from the portable device 102 is written. In this case, since the reference rate obtained in S1307 is 0, the process branches to S1309 without calculating the reference rate. That is, the first appearing word whose appearance count is 1 is treated as an unknown word to be included in the word list displayed on the portable device 102.

  On the other hand, if it is a registered word, the number of appearances is increased by one for the registered word (S1306). However, when the number of appearances in the text is stored for each word, the number may be added to the number of appearances. Thereafter, a value (reference rate) for determining whether or not to add the appearance word to the word list to be displayed on the portable device is calculated. This value is a value obtained by storing the reference count referenced when displayed on the portable device 102 in the reference rate table and dividing the reference count by the appearance count. The value of the reference rate is used as a reference value (S1307). The reference in the present embodiment means that when a list of appearance words is displayed on a portable device, the user selects from the list to display the meaning of a word. Therefore, even if it is determined to be an appearance word and displayed in a list, it is not referred to unless the user needs to know its meaning again, so the number of references does not increase. It is determined whether or not the reference value calculated in S1307 is greater than or equal to a predetermined reference value (that is, a predetermined value or more) (S1308). Word list data is created so as to display words and their meanings on the portable device 102 (S1309). On the other hand, if the reference value is smaller than the reference value, it is determined that it is not an unknown word and display is unnecessary, and even a meaningful word is not included in the word list. Therefore, it is not displayed on the portable device 102.

  In addition, for the referenced word from the word list displayed on the portable device, information related to the referenced word is transmitted to the document analysis apparatus 103. The document analysis apparatus 103 reconstructs the word list in a format for displaying the meaning of the referred word, for example, a format as shown in FIG. 8, and transmits it to the portable device 102 for display. At that time, the document analysis apparatus 103 adds 1 to the reference count of the referred word. In the present embodiment, even when a single word list is referred to a plurality of times, it is recorded as a reference count. For example, if the reference is made twice, the reference count increases by two. This is because it can be estimated that the necessity of displaying the word is high if the number of times of reference is large. On the other hand, one word may be counted once even if one word is referred to in one word list.

  In the procedure of FIG. 13, steps S1301 to S1306 are repeatedly executed for each word while paying attention to each of the appearing words, and after completion of registration or update of the number of appearances for all the appearing words, from step S1307 S1309 may be executed. In this case, since the reference value of a word newly registered in the reference rate table is 0, it is necessary to add to the word list regardless of the reference value only for the word whose appearance count is 1.

  The reference value in S1308 can be determined in consideration of, for example, how many times a certain word should be added to the word list after being referred to at the first appearance. For example, when a certain word is referenced at the first appearance, the number of appearances is 2, the number of references is 1, and the reference rate is 0.5 at the second appearance. Therefore, if the reference value is set to 0.5, the reference rate becomes equal to or higher than the reference value, and at this time, it is added to the word list. If it is not referenced the second time, the reference rate will be 1/3 when it appears for the third time, and will be smaller than the standard value 0.5, so it will no longer be added to the word list. In this way, by setting the reference value to, for example, 1 / N (however, it is rounded up in decimal representation), when it is referred to for the first time, it can be added to the word list until the Nth appearance.

  With the above-described configuration and procedure, the portable device 102 has the meanings of the first appearance word and the unknown word obtained by removing the noise word or the word having a low reference rate from the known words from the word group included in the target text. Can be displayed in a referable format. For this reason, the meaning of the word that will be unknown can be efficiently presented to the user. In the above procedure, only the appearance words stored in S1208 may be included in the word list as unknown words. In this case, it is not necessary to analyze the reference rate for known words.

[Modification]
In the procedure of FIG. 13, the word once excluded from the word list is never put in the word list again because the reference rate never increases. Therefore, for example, when the number of appearances reaches a predetermined number, 0 is set as the number of appearances and the reference number. By providing a mechanism for resetting the reference rate in this way, a word once dropped from the word list as a known word can be put on the word list again.

[Embodiment 2]
In Embodiment 1, there is a problem that when the learning period is short, the determination of unknown words is weak, that is, the accuracy of the determination is low. Therefore, it is possible to change a word that is determined not to be an unknown word according to the user's profile. This determination procedure will be described with reference to the flowchart of FIG. The procedure in FIG. 14 is executed when the access analysis unit 507 receives an analysis request from the portable device 102.

  The access analysis unit 507 newly adds a request source user of the analysis request to the access analysis storage unit 508 (S1401). At this time, it is determined whether or not users of the same tenant are registered in the reference rate table registered in the access analysis storage unit 508 (S1402). The same tenant user is a user who has received the same service in a multi-tenant environment. Specifically, it is another user having the same tenant ID. When there is a user of the same tenant, all the words appearing in the reference tenant registered in the reference rate table are referred to, and any user checks the word with the reference count of 0 (S1403). If any user has a word with a reference count of 0, the word is determined as a well-known word that is generally known within the same tenant, and the word ID, reference count, and appearance count are those of the registered user. And the record of the registered user is newly added to the reference rate table (S1404). If the same word is registered for a plurality of users, one of them is copied. That is, the reference rate required for users belonging to the same tenant as the request source user is also applied to the request source user.

  As a result, it becomes the same as the one having a learning period from the first time, and the known words are excluded from the unknown words and are not displayed.

[Embodiment 3]
In the third embodiment, an unknown word cannot be determined for each word unless learning is performed, and even an already known word may be displayed. Therefore, it is possible to change the display of words by classifying unknown words and specifying the display level.

An example of a UI for designating the level is shown in FIG. The word level is specified in advance for each dictionary. When the word dictionary presenting device 104 is composed of a plurality of units, the word dictionary presenting device 104 may have a different level.
A control 1501 is an example of a display control indicating a display level.

  By pressing the arrow control 1502, the displayed level is lowered and the display amount is increased. When the arrow control 1503 is pressed, the displayed level increases and the display amount decreases. For example, if the level of the word “wave equation” is 2, this word is displayed because it is higher than the display level “level 1” illustrated in FIG. When the display level is raised to 3 or more by pressing the control 1503, the word level of the “wave equation” becomes lower than the display level, so that it is not displayed in the word list.

  For example, when the control 1502 or 1503 is pressed while the word list is displayed on the portable device 102, information regarding the operation is transmitted to the document analysis apparatus 103. The document analysis apparatus 103 receives the information, selects from the word list a word given a new level or higher from the unknown word, reconstructs a new word list, and stores it in the portable device 102. Send and display. For this process, in the present embodiment, the process of FIG. 13 is executed regardless of the word level, and a maximum word list is created. However, the level specified in the dictionary for each word is included in the word list. The maximum word list is a word list that is displayed when the lowest display level is designated. When the display level is raised, a word having a level higher than the display level is selected from the maximum word list, and a new word list is reconstructed and transmitted to the portable device 102 for display.

  Thereby, words of a level lower than the designated display level are filtered, and the displayed words can be reduced.

[Other Examples]
The present invention can also be realized by executing the following processing. That is, software (program) that realizes the functions of the above-described embodiments is supplied to a system or apparatus via a network or various storage media, and a computer (or CPU, MPU, etc.) of the system or apparatus reads the program. It is a process to be executed.

Claims (9)

A word search device that determines an unknown word from words included in a text and outputs a word list of unknown words,
Word storage means for storing noise words and known words;
Search means for searching a dictionary for words that do not correspond to the noise word or the known word among the words included in the text;
A word registration unit for registering a word for which the search by the search unit has failed as the noise word, and newly registering a word for which the search has been successful as the known word;
Creating means for creating a word list including, as unknown words, words newly registered as the known words from the words included in the text;
Means for transmitting the word list to a terminal and displaying the word list.   The creation means includes, in the word list, words that are included in the text and that have a reference rate that is referred to in the word list that is greater than or equal to a predetermined value among the words that are stored as the known words as unknown words. The word search device according to claim 1, wherein:   The information regarding the word referred from the word list displayed on the said terminal is received, and ratio of the frequency | count referred to the frequency | count of appearance of the said word is calculated | required as the said reference rate. Word search device.   The word search apparatus according to claim 2, wherein the reference rate is obtained for each user who has transmitted the text analysis request.   5. The word search device according to claim 4, wherein when the text analysis request is received, a reference rate required for a user belonging to the same tenant as the request source user is also applied to the request source user. . The word has an associated level,
The word search device according to claim 1, wherein the creating unit does not include a word associated with a level lower than a designated level in the word list. The word search device according to any one of claims 1 to 6,
And a terminal for displaying a word list output from the word search device. A word search method for determining an unknown word from words included in a text and outputting a word list of unknown words,
A search step of searching a dictionary for a word that does not correspond to a noise word or a known word stored in the word storage means among words included in the text;
A word registration step of registering in the word storage means a word that has failed in the search in the search step as the noise word, and newly registering a word that has been successfully searched in the word storage means as the known word;
A creation step of creating a word list including, as unknown words, newly registered words as the known words from the words included in the text;
Transmitting the word list to a terminal and displaying the word list. A program for causing a computer to execute a word search method for determining an unknown word from words included in a text and outputting a word list of unknown words,
A search step of searching a dictionary for a word that does not correspond to a noise word or a known word stored in the word storage means among words included in the text;
A word registration step of registering in the word storage means a word that has failed in the search in the search step as the noise word, and newly registering a word that has been successfully searched in the word storage means as the known word;
A creation step of creating a word list including, as unknown words, newly registered words as the known words from the words included in the text;
A program for causing a computer to execute a step of transmitting and displaying the word list to a terminal.

Images