JP6916136B2 - Search support device, search support method, and search support program - Google Patents

Description

The present invention relates to a search support device or the like that assists a user in searching a document.

Conventionally, a search system for searching a document has been known. In the search system, when the user inputs a word or phrase included in a desired document to the search client as a search condition, the search server searches for the document containing the search condition and displays the search result on the search client.

For example, Patent Document 1 discloses a technique for storing user information and transmitting a search result in which an additional link is set on a web page based on the user information, as a technique for displaying a search result more suitable for the user's interest. Has been done.

Japanese Unexamined Patent Publication No. 2013-120406

For example, in a search system, the accuracy of the search result is affected by the words and phrases input to the user, so that the document desired by the user may not be displayed as the search result. In addition, even if the user's phrase is correct, if the phrase is used in a concept other than the user's intended concept, the document desired by the user may not be displayed as a search result, or the search result may be used. As a result, it may be difficult for the user to reach the document.

On the other hand, if the user can enter an appropriate phrase in the desired document as a search condition in as much detail as possible, the possibility of reaching the desired document is high. In this way, the phrase is entered in detail. This is often difficult for the user, and there is a problem that it takes time and effort.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a technique capable of easily and appropriately searching for a document desired by a user.

In order to achieve the above object, the search support device according to one viewpoint is a search support device that supports a search by a user, and includes a log collection unit that collects an operation log that can identify operation target data in the operation by the user. Candidate categories that specify one or more candidate categories based on the contents of the operation target data specified by the reception unit that accepts user conditions as search conditions from the user and one or more operation logs collected by the log collection unit. Specific part and candidate category A search including a search use category determination unit that determines one or more search use categories to be used for a search from one or more candidate categories specified by the specific part, user conditions, and a search use category. It is provided with a search control unit that executes a search depending on conditions.

According to the present invention, a document desired by a user can be easily and appropriately searched.

FIG. 1 is an overall configuration diagram of a search system according to an embodiment. FIG. 2 is a configuration diagram of an operation log table according to an embodiment. FIG. 3 is a configuration diagram of a search log table according to an embodiment. FIG. 4 is a flowchart of the learning process according to the embodiment. FIG. 5 is a flowchart of the operation log storage process according to the embodiment. FIG. 6 is a flowchart of the search process according to the embodiment. FIG. 7 is a flowchart of the category evaluation extraction process according to the embodiment. FIG. 8 is a flowchart of the search use category determination process according to the embodiment.

The embodiment will be described with reference to the drawings. It should be noted that the embodiments described below do not limit the invention according to the claims, and all of the elements and combinations thereof described in the embodiments are indispensable for the means for solving the invention. Is not always.

In the following description, the information may be described by the expression of "AAA table", but the information may be expressed by any data structure. That is, the "AAA table" can be called "AAA information" to show that the information does not depend on the data structure.

Further, in the following description, the process may be described with the "program" as the main body of operation, but the program is executed by a processor (for example, a CPU) to appropriately store the specified process as a storage resource (). Since it is performed using, for example, a memory) and / or a communication interface device (for example, a port), the main body of processing may be a processor. The process described with the program as the main body of operation may be a process performed by a device including a processor. It may also include hardware circuits that perform some or all of the processing performed by the processor. Computer programs may be installed on the device from the program source. The program source may be, for example, a program distribution server or a computer-readable storage medium.

FIG. 1 is an overall configuration diagram of a search system according to an embodiment.

The search system 1 includes one or more search clients 10, a search server 20, and a search log server 30. The search client 10, the search server 20, and the search log server 30 are connected via, for example, the network 40.

The network 40 is, for example, a network such as a wired LAN (Local Area Network) or a wireless LAN.

The search client 10 is an example of a search support device, and is configured by, for example, a PC (Personal Computer). The search client 10 includes a CPU (Central Processing Unit) 11, a memory 12, an input / output device 13, a communication circuit 14, and a storage device 15.

The communication circuit 14 is an interface such as a wired LAN card or a wireless LAN card, and communicates with other devices (for example, a search server 20 and a search log server 30) via a network 40.

The CPU 11 executes various processes according to a program stored in the memory 12 and / or the storage device 15.

The memory 12 is, for example, a RAM (RANDOM ACCESS MEMORY), and stores a program executed by the CPU 11 and necessary information.

The storage device 15 is, for example, a hard disk drive, an SSD (Solid State Drive), or the like, and stores a program executed by the CPU 11 and data used by the CPU 11. In the present embodiment, the storage device 15 stores the history search program 152 and the category evaluation program 153 as programs, and stores the operation log table 151 as data. Here, the history search program 152 and the category evaluation program 153 constitute a search support program.

The history search program 152 is a program that executes a process of managing the history of the operation log and a process of searching using the operation log. The category evaluation program 153 is a program that evaluates a category corresponding to a document and executes a process of determining an appropriate category. In the present embodiment, the category evaluation program 153 has a neural network model (hereinafter simply referred to as a model) created in advance for each category, and by inputting a plurality of words in the sentence for these models, this sentence Each category is evaluated, and the document category is specified based on the evaluation of each category. In the present embodiment, the category evaluation program 153 executes a process of learning a model corresponding to each category using the documents classified for each category as teacher data. Here, when the CPU 11 executes the history search program 152, a log collection unit, a reception unit, a search use category determination unit, a search control unit, and a search log acquisition unit are configured. Further, the CPU 11 executes the category evaluation program 153 to form a candidate category identification unit.

The operation log table 151 is a table that stores a history of a log (operation log) that can identify data used (referenced, etc.) by an operation by a user (user) of the search client 10. The details of the operation log table 151 will be described later.

The input / output device 13 includes, for example, an input device such as a mouse and a keyboard that accepts operation input by a user, and a display device that displays and outputs various information such as a display.

The search server 20 is composed of, for example, a PC. The search server 20 includes a CPU 21, a memory 22, a DVD-ROM drive 23, a communication circuit 24, and a storage device 25.

The communication circuit 24 is, for example, an interface such as a wired LAN card or a wireless LAN card, and communicates with other devices (for example, a search client 10 and a search log server 30) via a network 40.

The CPU 21 executes various processes according to a program stored in the memory 22 and / or the storage device 25.

The memory 22 is, for example, a RAM, and stores a program executed by the CPU 21 and necessary information.

The storage device 25 is, for example, a hard disk drive, an SSD, or the like, and stores a program executed by the CPU 21 and data used by the CPU 21. In the present embodiment, the storage device 25 stores the OS (Operating System) 251 and the search program 252 as programs, and stores the document data DB (database) 253 as data.

The search program 252 searches the document data corresponding to the search conditions transmitted from the search client 10 from the document data DB 253, and transmits the search result to the search source search client 10. The document data DB 253 stores the document data to be searched.

The DVD-ROM drive 23 is, for example, a drive for reading programs and data stored in a DVD-ROM, which is an example of a non-volatile recording medium.

The search log server 30 is composed of, for example, a PC. The search log server 30 includes a CPU 31, a memory 32, an input / output device 33, a communication circuit 34, and a storage device 35.

The communication circuit 34 is an interface such as a wired LAN card or a wireless LAN card, and communicates with other devices (for example, a search client 10 and a search server 20) via a network 40.

The CPU 31 executes various processes according to a program stored in the memory 32 and / or the storage device 35.

The memory 32 is, for example, a RAM, and stores a program executed by the CPU 31 and necessary information.

The storage device 35 is, for example, a hard disk drive, an SSD, or the like, and stores a program executed by the CPU 31 and data used by the CPU 31. In the present embodiment, the storage device 35 stores the search log program 352 as a program and stores the search log table 351 as data.

The search log program 352 stores the log (search log) related to the search by the user sent from the search client 10 or the search server 20 in the search log table 351 and searches for the search log request from the search client 10. It is a program that executes a process of acquiring a search log from a log table 351 and transmitting the search log. The search log table 351 is a table that stores search logs by a plurality of search clients 10. The details of the search log table 351 will be described later.

The input / output device 33 includes, for example, an input device that accepts operation input by a user such as a mouse and a keyboard, and a display device that displays and outputs various information such as a display.

Next, the detailed configuration of the operation log table 151 will be described.

FIG. 2 is a configuration diagram of an operation log table according to an embodiment.

The operation log table 151 is a table that manages logs (operation logs) related to user operations. The operation log (entry) of the operation log table 151 includes fields of a date and time 151a, a type 151b, a title 151c, and a reference destination 151d. The date and time 151a stores the date and time when the operation was performed. The type 151b stores a type indicating a user's operation target. Types include "document", "voice", "program", "site" and the like. "Document" is set when the user is using (referring to) a document file such as a Word (registered trademark) file, an Excel (registered trademark) file, or a text file, and "voice" is set. It is set when the user is referencing (viewing) audio data, and the "program" is set when the user uses a program file (for example, the identifiers are ".java", ".c", ".cpp"), etc. It is set when (editing, etc.) is being performed, and "site" is set when the user is browsing the site (web page of the site). The title to be operated is stored in the title 151c. The position (reference destination) of the data to be operated is stored in the reference destination 151d. Specifically, the reference destination 151d stores the directory and file name of the data to be operated and the URL (Uniform Resource Locator) of the site to be operated. The operation log table 151 may store a predetermined number of operation logs from the latest operation log, or store the latest operation log within a predetermined time range (for example, within 30 minutes). You may do so.

Next, the detailed configuration of the search log table 351 will be described.

FIG. 3 is a configuration diagram of a search log table according to an embodiment.

The search log table 351 is a table that manages logs (search logs) related to searches of a plurality of users. The search log (entry) in the search log table 351 includes fields for the date and time 351a and the search category 351b. The date and time when the search was performed is stored in the date and time 351a. The search category 351b stores a category added and used by the search client 10 in the search. The search log table 351 may store a predetermined number of search logs from the latest search log, or store the latest search log within a predetermined time range (for example, within 30 minutes). You may do so.

Next, the processing operation in the search system 1 will be described.

FIG. 4 is a flowchart of the learning process according to the embodiment.

The learning process is a process in which the category evaluation program 153 of the search client 10 (to be exact, the CPU 11 that executes the category evaluation program 153) learns the model of each category, and is executed before the search process described later is executed. NS. If the trained model is already prepared in the category evaluation program 153, the learning process may not be performed.

First, the category evaluation program 153 retrieves the document data (teacher data) divided into categories for learning from the storage device 15 (step S11). The teacher data may be acquired from another device. For example, when training a model for evaluating the category of "agriculture", a document "cultivation of mandarin oranges" that should be evaluated as the category of "agriculture" is prepared as teacher data.

Next, the category evaluation program 153 divides the extracted document content into words (step S12), and sorts the words into categories (step S13).

Next, the category evaluation program 153 executes the processing of loop A (step S14) for each category. Here, the category to be processed is referred to as a target category.

In the loop A, the category evaluation program 153 learns a model for evaluating the target category by using the words classified into the target category (step S14).

Then, when the processing of the loop A is completed for all the categories, the category evaluation program 153 ends the loop A and the learning result of each model that evaluates each category created in the loop A (each). The set value such as the weight of the model) is stored in the storage device 15 (step S15), and the learning process is completed.

According to this learning process, a plurality of models (specifically, model setting values) for evaluating each category are stored in the storage device 15, and the category evaluation program 153 uses each model. , You will be able to identify which category the content of the document belongs to.

Next, the operation log storage process will be described.

FIG. 5 is a flowchart of the operation log storage process according to the embodiment.

The operation log storage process is executed sequentially, for example, when the search client 10 is running.

The history search program 152 of the search client 10 accepts an operation by the user via the input / output device 13 (step S21). Next, the history search program 152 creates an operation log corresponding to the content of the received operation, stores it in the operation log table 151 (step S22), and ends the operation log storage process. Specifically, the history search program 152 creates an operation log including the date and time of the operation, the type of the operation target, the title of the data of the operation target, and the position of the data of the operation target, and stores the operation log in the operation log table 151.

Next, the search process will be described.

FIG. 6 is a flowchart of the search process according to the embodiment.

The search process is started when the history search program 152 receives the input of the search condition (user condition) from the user via the input / output device 13 in the search client 10.

The history search program 152 repeatedly executes the loop B processing (steps S31 to S43) for the number of operation logs (1 or more) to be processed in advance.

In loop B, first, the history search program 152 acquires an unprocessed operation log (target operation log) in the operation log to be processed from the operation log table 151 (step S31). Here, the operation log to be processed may be a predetermined number from the latest operation log. Next, the history search program 152 determines whether or not the operation indicated by the target operation log refers to the document by referring to the value of the target operation log type 151b (step S32).

As a result of this determination, when the operation indicated by the target operation log refers to the document (step S32: YES), the history search program 152 identifies the reference destination from the reference destination 151d of the target operation log, and refers to the reference destination. The contents of the document are acquired, the category evaluation extraction process (see FIG. 7) is executed for the acquired document (step S34), the category extracted by the category evaluation extraction process is specified as a candidate category, and the category is saved in the memory 12. (Step S43).

On the other hand, as a result of the determination in step S32, when the operation indicated by the target operation log does not refer to the document (step S32: NO), the history search program 152 uses a program file as the target of the operation indicated by the target operation log. It is determined whether or not there is (step S35). Specifically, the history search program 152 indicates that the identifier of the file to be operated is ".Java" indicating that it is a Java (registered trademark) language program file, and ". Java" indicating that it is a C language program file. If it is "c" or ".cpp" indicating that it is a C ++ program file, it is determined that the operation target is a program-type file.

As a result, when the target of the operation indicated by the target operation log is a program file (step S35: YES), the history search program 152 determines the programming language corresponding to the file identifier as a category (step S36). ), The determined category is saved in the memory 12 as a candidate category (step S43). For example, when the identifier is ".Java", the history search program 152 determines the category as Java, and when the identifier is ".c", the category is determined as C language, and the identifier is ".cpp". In the case of ", the category is determined as C ++.

On the other hand, as a result of the determination in step S35, when the target of the operation indicated by the target operation log is not a program file (step S35: NO), the history search program 152 uses the voice system as the target of the operation indicated by the target operation log. It is determined whether or not the file is (step S37).

As a result, when the target of the operation indicated by the target operation log is a voice file (step S37: YES), the history search program 152 identifies the reference destination from the reference destination 151d of the target operation log and refers to the reference destination. The data of the audio file of (step S38) is converted into text (step S38), the category evaluation extraction process (see FIG. 7) is executed for the text-converted document (step S39), and the category extracted by the category evaluation extraction process is used as a candidate category. It is saved in the memory 12 (step S43).

On the other hand, as a result of the determination in step S37, when the target of the operation indicated by the target operation log is not a voice file (step S37: NO), the history search program 152 determines whether the operation indicated by the target operation log is browsing the site. Whether or not it is determined (step S40).

As a result, when the operation indicated by the target operation log is browsing the site (step S40: YES), the history search program 152 identifies the reference destination from the reference destination 151d of the target operation log, and refers to the referenced site. The content (document) is acquired (step S41), the category evaluation extraction process (see FIG. 7) is executed for the acquired document (step S42), and the category extracted by the category evaluation extraction process is set as a candidate category in the memory 12 Save to (step S43).

On the other hand, as a result of the determination in step S40, if the operation indicated by the target operation log is not browsing the site (step S40: NO), the history search program 152 advances the process to step S43 and stores nothing as a candidate category. do not.

Then, when the processing of the loop B is completed for all the operation logs targeted for processing, the history search program 152 ends the loop B and determines the category (search use category) to be used for the search. The used category determination process (see FIG. 8) is executed (step S44), and the search is executed by the search condition in which the category determined by the search used category determination process is added to the user condition, that is, the search condition is sent to the search server 20. The transmission (step S45) is performed, and the process is terminated. As a result, the search result corresponding to the search condition is returned from the search server 20, and the history search program 152 displays the search result via the input / output device 13. In the present embodiment, when the history search program 152 transmits the search condition to the search server 20, the search use category included in the search condition is transmitted to the search log program 352 of the search log server 30.

Next, the category evaluation extraction process will be described.

FIG. 7 is a flowchart of the category evaluation extraction process according to the embodiment.

The category evaluation extraction process is a process corresponding to steps S34, S39, and S42 in the search process of FIG.

First, the category evaluation program 153 divides the content of the document (target document) to be classified into words into words (step S51), and repeatedly executes the loop C process (step S52) for the number of categories to be evaluated.

In the loop C, the category evaluation program 153 inputs the word of the target document into the trained model used for the evaluation of one of the categories to be processed (target category) in the category to be evaluated. The target category of the document is evaluated (step S52).

Then, when the processing of the loop C is completed for all the categories, the category evaluation program 153 ends the loop C and extracts the category having the highest evaluation among the plurality of categories (step S53). , End the category evaluation extraction process.

According to this category evaluation extraction process, the category of the target document can be appropriately extracted.

Next, the search use category determination process will be described.

FIG. 8 is a flowchart of the search use category determination process according to the embodiment.

The search use category determination process is a process corresponding to step S44 in the search process of FIG.

The history search program 152 acquires the categories used in the search by the plurality of search clients 10 from the search log server 30 (step S61). Here, as the category to be acquired from the search log server 30, a predetermined number (1 or more) of recent search log categories may be acquired, or may be a higher-ranked category that is frequently used within a predetermined time. ..

Next, the history search program 152 is used for the search based on the candidate category determined based on the operation log (the category stored in the memory 12 in step S43) and the category acquired from the search log server 30. The category (search use category) is determined (step S62). Here, as a method for determining the search use category, any method may be used. For example, the most used category (high frequency use category) is specified from the latest plurality of search logs, and the search use category is included in the candidate category. , If there is a match with the frequently used category, this candidate category may be decided as the search use category, and the number is the largest among the multiple recent search log categories and the candidate categories. The category may be determined as the search use category.

As described above, according to the search system 1 according to the present embodiment, it is possible to add a category based on the user's operation log in addition to the search condition input by the user to search, so that the user has trouble. A broader and more appropriate category can be used for the search without being overwhelmed, increasing the likelihood that the user's desired document can be effectively searched. In addition, since the category used for the search is determined using the category used in the past by other users, the search is appropriately performed not only according to the user to be searched but also according to the search tendency of the past user. Therefore, the document desired by the user can be searched more effectively.

The present invention is not limited to the above-described embodiment, and can be appropriately modified and implemented without departing from the spirit of the present invention.

For example, in the above embodiment, a model is used to evaluate a plurality of categories of a document and the categories corresponding to the documents are extracted. However, the method of extracting the categories of a document uses a model. The method is not limited to this, and any other method may be used.

Further, in the above embodiment, the search log server 30 manages the search log in association with the user who performed the search, and in step S61, another user (a user different from the user who is searching) manages the search log. You may want to retrieve only the search log categories used in the search, which belong to other users (eg, the same group (company, organization, department, class, etc.)) that have a given relationship with the user you are searching for. You may want to get only the search log category of other users). In this way, it is possible to search using the category used by another user who has a high tendency to perform a search common to the user, so that the possibility that the user can search the desired document is higher.

1 ... Search system, 10 ... Search client, 11 ... CPU, 12 ... Memory, 13 ... I / O device, 15 ... Storage device, 20 ... Search server, 30 ... Search log server, 151 operation log table, 152 ... History search program , 153 ... Category evaluation program

Claims (9)

A search support device that supports searches by users.
A log collection unit that collects operation logs that can identify operation target data in the operation by the user,
A reception unit that accepts user conditions as search conditions from the user,
A candidate category identification unit that specifies one or more candidate categories based on the contents of operation target data identified by one or more operation logs collected by the log collection unit, and a candidate category identification unit.
A search use category determination unit that determines one or more search use categories to be used for a search from one or more of the candidate categories specified by the candidate category identification unit.
A search control unit that executes a search according to a search condition including the user condition and the search use category.
A search support device equipped with. The search support device according to claim 1, wherein the candidate category specifying unit specifies one or more categories corresponding to the contents of the operation target data specified by the latest plurality of operation logs as candidate categories. The search support device according to claim 1 or 2, wherein the candidate category specifying unit identifies the candidate category based on the content of the document data when the operation target data is document data. The search support device according to any one of claims 1 to 3, wherein the candidate category specifying unit specifies the programming language in the candidate category when the operation target data is programming language data. .. The candidate category specifying unit is described in any one of claims 1 to 4, which specifies the candidate category based on the content of the page of the site when the operation target data is a page of the site. Search support device. When the operation target data is voice data, the candidate category specifying unit converts the voice data into text and specifies the candidate category based on the textualized content. Any of claims 1 to 5. The search support device described in item 1. It also has a search log acquisition unit that acquires one or more search logs that can identify the category used by the search.
The search use category determination unit searches for one or more candidate categories specified by the candidate category identification unit and categories specified by one or more search logs acquired by the search log acquisition unit. The search support device according to any one of claims 1 to 6, which determines the use category. It is a search support method using a search support device that supports searches by users.
A log collection step for collecting operation logs that can identify operation target data in the operation by the user, and
A reception step that accepts user conditions as search conditions from the user, and
A candidate category identification step that identifies one or more candidate categories based on the content of operation target data identified by one or more operation logs collected in the log collection step, and a candidate category identification step.
A search use category determination step for determining one or more search use categories to be used for a search from one or more of the candidate categories identified in the candidate category identification step, and a search use category determination step.
A search step for executing a search according to a search condition including the user condition and the search use category.
Search support method with. It is a search support program to be executed by the computers that make up the search support device that supports the search by the user.
The computer
A log collection unit that collects operation logs that can identify operation target data in the operation by the user,
A reception unit that accepts user conditions as search conditions from the user,
A candidate category identification unit that specifies one or more candidate categories based on the contents of operation target data identified by one or more operation logs collected by the log collection unit, and a candidate category identification unit.
A search use category determination unit that determines one or more search use categories to be used for a search from one or more of the candidate categories specified by the candidate category identification unit.
A search control unit that executes a search according to a search condition including the user condition and the search use category.
A search support program that works.

Images