JP7396190B2 - Extraction program, extraction method and extraction device - Google Patents

Description

The present invention relates to an extraction program, an extraction method, and an extraction device.

2. Description of the Related Art Conventionally, a technique is known in which content is searched using words obtained by segmenting a search keyword through morphological analysis.

Japanese Patent Application Publication No. 2015-138351

However, the conventional technology has a problem in that search keywords may not be separated appropriately.

For example, in conventional technology, morphological analysis is performed based on words registered in a dictionary. Here, consider a case where the word "Fujitsu Ltd." is registered in the dictionary, but the word "Fujitsu" is not registered. At this time, if the user enters the search keyword "Fujitsu" and performs a search, it may not be possible to search for a document in which the word "Fujitsu Ltd." is written. It is assumed that the user inputs the search keyword ``Fujitsu,'' which is an abbreviation, with the intention that a document containing the word ``Fujitsu Ltd.'' will be included in the search results.

One aspect aims to appropriately segment search keywords.

In one aspect, the extraction program performs a process of identifying the search ranking of correct answer knowledge in search results when a search is performed using a search keyword after separation for each of a plurality of separation patterns for a search keyword before separation using data. Let the computer run it. The data is data showing the correspondence between the search keyword before the break and the correct answer knowledge for the search keyword before the break. The extraction program causes the computer to execute a process of extracting, from among the plurality of delimitation patterns, a delimitation pattern whose specified search ranking satisfies a predetermined condition as a delimitation pattern to be applied to the search keyword before delimitation.

In one aspect, search keywords can be appropriately separated.

FIG. 1 is a diagram showing an example of the configuration of an extraction device. FIG. 2 is a diagram showing an example of the data structure of dictionary information. FIG. 3 is a diagram illustrating extraction of a delimiter pattern. FIG. 4 is a flowchart showing the flow of the extraction process. FIG. 5 is a diagram illustrating an example of a hardware configuration.

Embodiments of the extraction program, extraction method, and extraction device according to the present invention will be described in detail below based on the drawings. Note that the present invention is not limited to this example. Moreover, each embodiment can be combined as appropriate within a consistent range.

The configuration of an extraction device according to an embodiment will be explained using FIG. 1. FIG. 1 is a diagram showing an example of the configuration of an extraction device. The extraction device 10 receives input of a search keyword from a user. Further, the extraction device 10 outputs data in a dictionary format in which delimitation patterns of search keywords are registered.

As shown in FIG. 1, the extraction device 10 includes an input section 11, an output section 12, a storage section 13, and a control section 14. The input unit 11 is an interface for inputting data. For example, the input unit 11 receives data input via input devices such as a mouse and a keyboard. Further, the output unit 12 is an interface for outputting data. For example, the output unit 12 outputs data to an output device such as a display.

The storage unit 13 is an example of a storage device that stores data, programs executed by the control unit 14, and the like, and is, for example, a hard disk or a memory. The storage unit 13 includes correct answer knowledge information 131, dictionary information 132, and content DB 133.

The correct knowledge information 131 is the rank (rank_base) of the correct knowledge. For example, the correct answer knowledge is the one that most matches the user's intention among the search results for the search keyword input by the user.

For example, when a user enters the search keyword "input/output design document" to search for a desired content from a group of contents, the content desired by the user is located third in the search results. Suppose that In this case, the correct knowledge information 131 is information indicating that the desired content was ranked third in the search ranking for the search keyword "input/output design document."

The dictionary information 132 is information for performing morphological analysis. FIG. 2 is a diagram showing an example of the data structure of dictionary information. As shown in FIG. 2, the dictionary information 132 includes a search keyword before separation and a search keyword after separation. In the example of FIG. 2, the search keyword before the break is "input/output design document" and the search keyword after the break is "[input/output], [design document]".

The content DB 133 stores content. The contents include, for example, document files, text files, spreadsheet files, and the like.

Returning to FIG. 1, the control unit 14 uses, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), etc., to allow programs stored in an internal storage device to use the RAM as a work area. This is achieved by executing as . Further, the control unit 14 may be realized by, for example, an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). The control unit 14 includes an analysis unit 141, a search unit 142, a specification unit 143, and an extraction unit 145.

The analysis unit 141 performs morphological analysis on the search keyword before separation. For example, the search keyword before the break is input by the user via the input unit 11. The analysis unit 141 uses morphological analysis to decompose the search keyword before delimitation into elements based on a predetermined delimitation pattern. Note that in this example, an element is assumed to be one morpheme or a character string formed by combining a plurality of morphemes.

The search unit 142 searches for content using a search keyword before division or a search keyword after division. The search unit 142 searches for content stored in the content DB 133. The search unit 142 also ranks the content obtained as a search result. For example, the search unit 142 assigns a higher ranking in descending order of similarity to the search keyword.

Using data, the identification unit 143 identifies the search ranking of correct knowledge in the search results when a search is performed using the search keyword after separation, for each of a plurality of separation patterns for the search keyword before separation. The data is data showing the correspondence between the search keyword before the break and the correct answer knowledge for the search keyword before the break.

The extraction unit 144 extracts, from among the plurality of break patterns, a break pattern whose specified search ranking satisfies a predetermined condition as a break pattern to be applied to the search keyword before break. Further, the extraction unit 144 registers information obtained by dividing the search keyword before division using the extracted division pattern in the storage unit 13 as data in a dictionary format, that is, dictionary information 132.

FIG. 3 is a diagram illustrating extraction of a delimiter pattern. First, as shown in FIG. 3, the analysis unit 141 performs morphological analysis on the search keyword before separation using a plurality of separation patterns. Here, it is assumed that the search keywords before separation are word w_1+word w_2+word w_3+word w_4. Note that "+" is an operator representing a combination of character strings. For example, the separation pattern includes a pattern that separates all of word w_1, word w_2, word w_3, and word w_4, a pattern that leaves word w_1 and word w_2 combined, and the like.

At this time, the search unit 142 searches for content using each of the post-delimitation search keywords corresponding to each delimitation pattern. The specifying unit 143 specifies the delimiter pattern with the highest search accuracy.

The extracting unit 144 registers the delimiter pattern specified by the specifying unit 143 as dictionary information 132. For example, the extraction unit 144 selects, among the plurality of break patterns, a break pattern whose specified search ranking is higher than the search ranking of the correct answer knowledge in the search results when a search is performed using the search keyword before the break. Extract as a separation pattern to be applied to the search keyword before separation.

This will be explained by giving a more specific example. It is assumed that the user desires to obtain a file called "input/output screen design document.xlsx" as a search result, and inputs the search keyword "input/output design document." At this time, it is assumed that when the search keyword "input/output design document" is used as is, the ranking of "input/output screen design document.xlsx" in the search results is third. Further, it is assumed that the search keyword "input/output design document" and the information that "input/output screen design document.xlsx" was ranked third are stored as correct answer knowledge information 131.

First, the analysis unit 141 divides the "input/output design document" into four patterns: "input/output design document," "input/output design document," "input/output design document," and "input/output design document." Disassemble. Note that a space is a delimiter and represents an AND condition in a search.

Then, the search unit 142 performs a search using each pattern of search keywords. It is assumed that when the search keyword is "input/output design document", the ranking of "input/output screen design document.xlsx" in the search results is 10th place. Further, it is assumed that when the search keyword is "input/output design document", the ranking of "input/output screen design document.xlsx" in the search results is 6th place. It is also assumed that when the search keyword is "input/output design document," the ranking of "input/output screen design document.xlsx" in the search results is 3rd place. Further, it is assumed that when the search keyword is "input/output design document", the ranking of "input/output screen design document.xlsx" in the search results is second.

At this time, the specifying unit 143 specifies the delimiter pattern corresponding to the "input/output design document" having the highest rank. Then, the extraction unit 144 expresses the delimiter pattern as “[input/output], [design document]” and adds it to the dictionary information 132. Note that the expression format of the delimiter pattern is not limited to the above.

Here, the analysis unit 141 performs morphological analysis on search keywords registered in the dictionary information 132 according to the dictionary information 132. In other words, as shown in FIG. 2, if the search keyword after the break that corresponds to the search keyword "input/output design document" before the break has been added to the dictionary information 132, the analysis unit 141 changes the "input/output design document" to "input/output design document". It is broken down into "input/output design document". Then, the search unit 142 performs a search using the search keyword "input/output design document."

FIG. 4 is a flowchart showing the flow of the extraction process. As shown in FIG. 4, the analysis unit 141 first decomposes a search keyword into elements by morphological analysis (step S101). Next, the search unit 142 searches for each delimiter pattern in the morphological analysis (step S102). The specifying unit 143 then specifies the ranking of the correct knowledge in the search results (step S103). The extraction unit 144 extracts the delimiter pattern with the highest rank and registers it in the dictionary (step S104).

As described above, the specifying unit 143 uses data to specify, for each of a plurality of break patterns for the search keyword before break, the search ranking of the correct answer knowledge in the search results when a search is performed using the search keyword after break. do. The data is data showing the correspondence between the search keyword before the break and the correct answer knowledge for the search keyword before the break. The extraction unit 144 extracts, from among the plurality of break patterns, a break pattern whose specified search ranking satisfies a predetermined condition as a break pattern to be applied to the search keyword before break. In this way, the extraction device 10 can extract delimiter patterns that actually have good search accuracy. Therefore, according to this embodiment, search keywords can be appropriately divided.

The extraction unit 144 selects, from among the plurality of delimitation patterns, a delimitation pattern whose identified search ranking is higher than the search ranking of the correct answer knowledge in the search results when searching with the search keyword before the delimitation. Extract as a delimiter pattern to be applied to search keywords. Thereby, it is possible to extract only the break pattern that improves the search accuracy compared to the search keyword before the break.

The extraction unit 144 registers information obtained by dividing the search keyword before division using the extracted division pattern into the storage unit as data in a dictionary format. In this way, by storing delimiter patterns that improve search accuracy as a dictionary, highly accurate searches can be performed.

Information including processing procedures, control procedures, specific names, and various data and parameters shown in the above documents and drawings can be changed arbitrarily unless otherwise specified. Furthermore, the specific examples, distributions, numerical values, etc. described in the examples are merely examples, and can be changed arbitrarily.

Furthermore, each component of each device shown in the drawings is functionally conceptual, and does not necessarily need to be physically configured as shown in the drawings. That is, the specific form of distributing and integrating each device is not limited to what is shown in the drawings. In other words, all or part of them can be functionally or physically distributed and integrated into arbitrary units depending on various loads, usage conditions, and the like. Furthermore, all or any part of each processing function performed by each device can be realized by a CPU and a program that is analyzed and executed by the CPU, or can be realized as hardware using wired logic.

FIG. 5 is a diagram illustrating an example of a hardware configuration. As shown in FIG. 5, the extraction device 10 includes a communication interface 10a, an HDD (Hard Disk Drive) 10b, a memory 10c, and a processor 10d. Further, each part shown in FIG. 5 is interconnected by a bus or the like.

The communication interface 10a is a network interface card or the like, and communicates with other servers. The HDD 10b stores programs and DB that operate the functions shown in FIG.

The processor 10d reads a program that executes the same processing as each processing unit shown in FIG. It is a hardware circuit. That is, this process executes the same functions as each processing unit included in the extraction device 10. Specifically, the processor 10d reads a program having the same functions as the analysis section 141, the search section 142, the identification section 143, and the extraction section 145 from the HDD 10b or the like. The processor 10d then executes a process that performs the same processing as the analysis unit 141, search unit 142, identification unit 143, extraction unit 145, and the like.

In this way, the extraction device 10 operates as an information processing device that executes a learning method by reading and executing a program. Further, the extraction device 10 can also realize the same functions as in the above-described embodiments by reading the program from the recording medium using a medium reading device and executing the read program. Note that the programs in other embodiments are not limited to being executed by the extraction device 10. For example, the present invention can be similarly applied to cases where another computer or server executes a program, or where these computers or servers cooperate to execute a program.

This program can be distributed via a network such as the Internet. In addition, this program is recorded on a computer-readable recording medium such as a hard disk, flexible disk (FD), CD-ROM, MO (Magneto-Optical disk), or DVD (Digital Versatile Disc), and is read from the recording medium by the computer. It can be executed by being read.

10 Extraction device 11 Input section 12 Output section 13 Storage section 14 Control section 131 Correct answer knowledge information 132 Dictionary information 133 Content DB
141 Analysis Department,
142 Search section 143 Specification section 144 Extraction section

Claims (5)

Using data showing the correspondence between the search keyword before the break and the correct answer knowledge for the search keyword before the break, a search is performed using the search keyword after the break for each of the plurality of break patterns for the search keyword before the break. identify the search ranking of the correct answer knowledge in the search results when
An extraction program that causes a computer to perform a process of extracting, among the plurality of break patterns, a break pattern whose search ranking satisfies a predetermined condition as a break pattern to be applied to the search keyword before the break. . The process of extracting includes, among the plurality of break patterns, a break in which the identified search ranking is higher than the search ranking of the correct answer knowledge in the search results when a search is performed using the search keyword before the break. 2. The extraction program according to claim 1, wherein the pattern is extracted as a break pattern to be applied to the search keyword before the break. 2. The method further comprises performing a process of registering information obtained by delimiting the search keyword before the delimitation in the storage unit as data in a dictionary format using the delimitation pattern extracted by the extraction process. extraction program. Using data showing the correspondence between the search keyword before the break and the correct answer knowledge for the search keyword before the break, a search is performed using the search keyword after the break for each of the plurality of break patterns for the search keyword before the break. identify the search ranking of the correct answer knowledge in the search results when
An extraction method characterized in that a computer performs a process of extracting, from among the plurality of break patterns, a break pattern whose specified search ranking satisfies a predetermined condition as a break pattern to be applied to the search keyword before the break. . Using data showing the correspondence between the search keyword before the break and the correct answer knowledge for the search keyword before the break, a search is performed using the search keyword after the break for each of the plurality of break patterns for the search keyword before the break. a specifying unit that specifies the search ranking of the correct answer knowledge in the search results when
an extraction unit that extracts, from among the plurality of break patterns, a break pattern in which the specified search ranking satisfies a predetermined condition as a break pattern to be applied to the search keyword before the break;
An extraction device characterized by having:

Images