JP6916437B2 - Information processing device, its control method, and program - Google Patents

Description

The present invention relates to an information processing device capable of selecting a character string to be corrected with respect to an erroneous character string included in a search condition , a control method thereof, and a programming technique.

In recent years, personal computers have become widespread in both companies and homes, and the Internet has become familiar, so there are many opportunities to enter character strings from the keyboard.

However, it takes a lot of skill to master the keyboard, and even a skilled person often makes an erroneous input (spelling mistake).

Also, in recent years, with voice recognition, a technology for converting a human voice into a character string by computer software without using a keyboard has become widely used. However, since the accuracy of voice recognition is not 100%, spelling mistakes occur. And various techniques have been developed to deal with spelling mistakes efficiently.
In Patent Document 1, an electronically stored dictionary is prepared in advance, a misrecognized part of the character string received by voice recognition is specified, and further, the misrecognized character string from the dictionary is dealt with. The technique for calculating the similarity between the correction candidate character string and the erroneously recognized character string and the correction candidate character string is described.

Further, in Patent Document 2, in a search system, a technique of storing a query (character string) input by a user in the past as a log and performing a spelling check based on the query log when a new user inputs a query (character string). Is described.

Japanese Unexamined Patent Publication No. 2012-063545 Japanese Unexamined Patent Publication No. 2005-267638

However, in Patent Document 1, it is necessary to prepare a dictionary in advance. The words contained in this dictionary must be generic and applicable to all disciplines. In that case, man-hours for preparing the dictionary in advance and maintenance man-hours for updating the dictionary such as adding new words are required. In addition, even if the character string newly input by the user is in a specific field, there are a large number of correction candidates in order to search for the character string of the correction candidate from the dictionary prepared for general purposes, and the similarity between the character strings is high. Even if the calculation of is calculated with high accuracy, there is a problem that irrelevant candidates are presented to the user with high priority.

Further, in Patent Document 2, although the query log of a specific user is stored and used for spelling check, it is used for general purposes regardless of the case where the user frequently uses only a specific query. In some cases, it takes a long time to collect so many logs for spelling check, and because of its versatility, irrelevant candidates should be presented to the user with high priority as in Patent Document 1. There is a problem of becoming.

An object of the present invention is to provide a technique capable of selecting a character string to be corrected for an erroneous character string included in a search condition.

In order to achieve the above object, the present invention is an information processing apparatus that acquires document data searched for a search condition including a character string from a plurality of document data, and an error is found in the search condition. A character that corrects the erroneous character string based on a specific means for identifying a certain character string, a character string included in the document data acquired for the search condition, and the specified erroneous character string. A selection means for selecting a column from a character string included in the acquired document data is provided, and the selection means is included in the character string included in the document data and the search condition in the acquired document data. It is characterized in that the character string to be corrected is selected based on the value related to the co-occurrence with the character string .

According to the present invention, it is possible to select a character string to be corrected for an erroneous character string included in the search condition.

It is a figure which shows an example of the functional structure which concerns on embodiment of this invention. It is a block diagram which shows an example of the hardware composition of the information processing apparatus which concerns on embodiment of this invention. It is a figure which shows an example of the flowchart of the search process which concerns on embodiment of this invention. It is a figure which shows an example of the flowchart of the processing of the spelling check which concerns on embodiment of this invention. It is a figure which shows an example of the document to be searched which concerns on embodiment of this invention. It is a figure for demonstrating an example of the extracted word storage part which concerns on embodiment of this invention, and the storage part about the information of the co-occurrence relationship between a word included in a search condition and an extracted word. It is a figure for showing an example of the candidate word list related to a spelling mistake in the word dictionary storage part which concerns on embodiment of this invention. It is a figure for demonstrating an example of the modification candidate storage part which concerns on embodiment of this invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a diagram showing an example of a functional configuration according to an embodiment of the present invention.

The input character string receiving unit 101 receives an input character string that is a search condition for searching a text document (corresponding to document data). In the present embodiment, the example of the document data is a text document, but the document data is not limited to the text document as long as the document is included, and the document data such as a PDF file may be used. When the information processing device in the present invention is a computer that functions as a client device, the input character string may be a functional unit that directly receives input from a user who uses the computer. Further, it may be a functional unit that receives from another application program or the like running on the computer. When the information processing device in the present embodiment is a server on a network, it may be a functional unit that receives communication information from another information processing device via the network.
The document search unit 102 searches for a text document stored in the document storage unit 121 based on the input character string acquired in the above procedure. The search may be performed on a pre-generated index from the text document rather than on the text document itself. The technique for searching and indexing text documents is a well-known technique, so details are omitted.

The word extraction unit 103 analyzes a character string included in the text document from the text document that is the result of the search by the document search unit 102, and extracts a word.

The dictionary search unit 104 receives input received by the input character string reception unit 101 from the word dictionary storage unit 122 (corresponding to the management means) that stores and manages word information (corresponding to correction data) prepared in advance. A part of a character string (substring), that is, an input character string is analyzed to search for a word based on a word (which may contain spelling mistakes). However, since it is not known in advance whether the character string of the word to be searched is correct or incorrect, it is not only the word that exactly matches the character string, but it is similar considering that there is a spelling error. It searches for a character string that is determined to be. Since dictionary search for spell checking is a well-known technique, details are omitted.

In the similarity determination unit 105, the word extracted by the word extraction unit 103 or the word searched by the dictionary search unit 104 is a substring of the input character string, and there is a possibility that the word (including a spelling error). Determine how similar it is to a character string). The result is calculated numerically.

The correction candidate storage unit 106 temporarily stores a word that seems to be similar to a word that is a substring of the input character string in the storage unit together with the degree of similarity. The modification candidate is presented to the user or an application using the system of the present invention by the modification candidate presentation unit 107, and the selection result of the user or the application is received by the modification result reception unit 108 and obtained from the input character string. Correct the word. The corrected result becomes a search condition again, and the document search unit 102 searches the document storage unit 121 again.

FIG. 2 is a block diagram showing an example of a hardware configuration of an information processing device (corresponding to an information processing server) included in the information processing system of the present invention.

As shown in FIG. 2, the information processing device 100 and the application server 140 have a CPU (Central Processing Unit) 201, a RAM (Random Access Memory) 202, a ROM (Read Only Memory) 203, and an input controller 205 via the system bus 204. , Video controller 206, memory controller 207, communication I / F controller 208 and the like are connected. The CPU 201 comprehensively controls each device and controller connected to the system bus 204.

Further, the ROM 203 or the external memory 211 will be described later, which is necessary for realizing the functions executed by the BIOS (Basic Input / Output System) and the OS (Operating System) which are the control programs of the CPU 201, and each server or each PC. Various programs etc. are stored. In addition, information necessary for carrying out the present invention is stored. The external memory may be a database.

The RAM 202 functions as a main memory, a work area, and the like of the CPU 201. The CPU 201 realizes various operations by loading a program or the like necessary for executing the process from the ROM 203 or the external memory 211 into the RAM 202 and executing the loaded program.

Further, the input controller 205 controls input from a pointing device such as a keyboard (KB) 209 or a mouse (not shown).

The video controller 206 controls the display on a display such as the display 210. The display may be a display such as a liquid crystal display. These are used by the administrator as needed.

The memory controller 207 is an external storage device (hard disk (HD)) for storing boot programs, various applications, font data, user files, edit files, various data, etc., a flexible disk (FD), or a PCMCIA (Personal Computer). It controls access to an external memory 211 (corresponding to a storage means) such as a compact flash (registered trademark) memory connected to a Memory Card International Association card slot via an adapter.

The communication I / F controller 208 connects and communicates with an external device via the network, and executes communication control processing on the network. For example, communication using TCP / IP (Transmission Control Protocol / Internet Protocol) is possible.

The CPU 201 can display the outline font on the display 210 by executing the outline font expansion (rasterization) process on the display information area in the RAM 202, for example. Further, the CPU 201 enables a user instruction by a mouse cursor (not shown) or the like on the display 210.

Various programs described later for realizing the present invention are recorded in the external memory 211, and are executed by the CPU 201 by being loaded into the RAM 202 as needed.
FIG. 3 is a diagram showing an example of a flowchart of a search process according to an embodiment of the present invention. Each step of the flowchart of FIG. 3 is executed by the CPU 201 on the information processing apparatus 100.

In step S301, the input character string receiving unit 101 accepts the input character string as a condition for searching the text document in the document search unit 102 (corresponding to the receiving means). As an example for explaining the present invention, it is assumed that the character string "artificial intelligence, machine learning, big day?" Is accepted. The last "?" Is a spelling error (corresponding to a correction character string) in the input of the word "big data", and is a correction target. For example, it is assumed that one character is insufficient for "Big Day", or an incorrect character is input and it is "Big Day".

In step S302, the document search unit 102 acquires (searches) a list of text documents that match the conditions from the document storage unit 121 (corresponding to the search means) using the above-mentioned input character string as a search condition. An example of the document storage unit 121 will be described with reference to FIG.

In FIG. 5, 501 is a text document to be searched. In the example of FIG. 5, the text documents are 501a to 501n. Among these text documents 501, the documents including the above-mentioned search conditions "artificial intelligence" and "machine learning" are the search results, but in the example of FIG. 5, there are four of 501a to 501d. Words that match the search criteria are underlined. For example, in the text document 501a, "artificial intelligence" and "machine learning" are the same.

The word "big data" is also included, but it is not considered a match because the search criteria spelled "Big Day?".

In step S303, the word extraction unit 103 analyzes the above-mentioned text document as a search result and extracts a word included in any of the text documents.

The extracted words are represented as a list, for example, in the extracted word storage unit 311 (vertical column from "big data" to "digital") in FIG. FIG. 6 will be described here.

FIG. 6 is a two-dimensional table of the list of the extracted word storage unit 311 and the portion (601 in the horizontal column) in which the words of the search condition are arranged for each document as described above. Weighting is described in the weighting column 602 where the extracted word storage unit 311 and each document 601 intersect.

For example, in the upper left column, since the text document 501 contains the word "big data", "2" is described in the column to the right of the extracted word "big data". This is because two words "artificial intelligence" and "machine learning" appear in the text document in the search condition, that is, "big data" is two words among the words included in the search condition. It shows that they are co-occurring.

This weighting is just an example. For example, in a search system, each word that appears in a text document is usually given a "weight". Words that appear many times in one text document have a higher weight, while words that appear many times in different text documents have a lower weight, and so on. In consideration of these weights, the value of the weight column 602 in FIG. 6 may be calculated.

Further, in FIG. 6, only two words of the search condition appear in each text document, but when there are more words, for example, four words A1, A2, B1 and B2, the words A1 and A2 co-occurrence frequency. The weight of the extracted words co-occurring with the words A1 and A2 is also high, while the weight of the extracted words co-occurring with the words B1 and B2 is low because the frequency of co-occurrence of the words B1 and B2 is low. And so on. Therefore, the method of calculating the value in the weighting column 602 is arbitrary and is a design matter. In the present invention, any calculation method is included.

Further, the extracted word storage unit 311 does not include the two words "artificial intelligence" and "machine learning", but this is because it is treated as if there was no spelling error in the explanation described later, and the step 303. If the determination is not made in, the extracted word storage unit 311 may be provisionally included.

Similarly, in the extracted word storage unit 311 of FIG. 6, all the words appearing in each text document except for "artificial intelligence" and "machine learning" are described, but "Big day?" In this example. As long as it is a character string that can be determined in advance that it is not even a candidate for the spelling check, it does not necessarily have to be included in the extracted word storage unit 311.

Return to the explanation of the flowchart. In step S304, a spelling check is performed on the words included in the input character string that is the search condition. The spelling check is a well-known technique. However, a part related to the features of the present invention will be described in detail with reference to the flowchart of FIG.

Steps S401 to S406 are processes that are repeated while paying attention to each word (including those with spelling mistakes) extracted from the input character string of the search condition.

In step S402, it is determined whether or not the same word as the word of interest is in the extracted word storage unit 311. For example, it is determined whether or not there is a "big day?" If there is a match as a character string (YES), the process proceeds to step S406, and the next focus shifts to the next word. If there is no match (NO), the process proceeds to step S403.

In step S403, the similarity determination unit 105 selects the word of interest from the words included in the text document of the search result stored in the extracted word storage unit 311 and the words included in the word dictionary storage unit 122. Select similar ones and calculate the similarity. The degree of similarity is calculated based on, for example, the degree of matching of two words as a character string. The spelling check and similarity calculation processes are well-known techniques and details are omitted. FIG. 7 will be described as an example of the word dictionary storage unit.

In FIG. 7, only words having a heading similar to "Big day?" Are displayed, but other words are actually registered. This is a state in which words having a similar heading (702) are listed for the word 701 of interest.

In addition, the problem of using a dictionary as in Patent Document 1 has been described above, but this is a case of preparing a general-purpose dictionary, for example, a company may develop a question answering system for a specific product. In such a case, if only the technical terms related to the product are registered, there are few words that are mistakenly similar to the word of interest, and the maintenance man-hours are small, so it is effective even if it is used. It is possible to prevent the problem of.

Further, an example of a list of words determined to be similar will be described with reference to FIG. FIG. 8 illustrates the correction candidate storage unit 312. The result of step S403 is stored in the correction candidate storage unit 312.

The correction candidate storage unit 312 stores the notation 802 of words determined to be similar to the spelling error candidate 801 and the degree of similarity 803 thereof. Further, the source 804 may be stored so that the user or the application can determine that the modification candidate acquired from the word dictionary storage unit 122 is not used. When the search system does not make a judgment by default and presents it to a user or an application, the information of the source 804 may be presented together with the correction candidate as a basis for the judgment.

In step S404, it is determined from the extracted word storage unit 311 and the word dictionary storage unit 122 (if necessary) whether or not a character string similar to the word of interest can be found. Specifically, a threshold value of similarity is set and stored in a storage unit such as a program code or a setting file. For example, if a threshold value of "0.5" is set and the similarity is lower than that value, it may be determined that the spelling error is not a correction candidate.

If it is determined that there is something similar to the word of interest (YES), the process proceeds to step S405. If it is determined that there is no similar thing (NO), the process proceeds to step S406, and the next focus shifts to the next word.

In step S405, the degree of similarity is changed for words that are considered to be similar to the word of interest.

An example will be given. Already, in FIG. 6, the degree of co-occurrence of the extracted word and the word included in the search condition (pointing to the number of times it occurs in a linguistic environment such as an utterance, a sentence, or a context in which the language is the same) is calculated (corresponding to the calculation means). ) Described the weighting table. For example, "big data" with a similarity of 0.8 co-occurs with a total of four search conditions in three text documents from the table in FIG. 6, so the similarity is quadrupled to "3.2". Since only one "bit data" having a similarity of 0.6 co-occurs, it may be changed to "0.6" or the like. This method of calculating the degree of co-occurrence is just an example, and any calculation method may be used. In the present invention, those arbitrary calculation methods are also included.

This completes the description of the flowchart of FIG. 4, and returns to the description of the flowchart of FIG.

In step S305, the correction candidate presentation unit 107 presents correction candidates for words in the search condition determined to have a spelling error. For example, out of the notation 802 in FIG. 8, only the ones described as “extracted” in the source 804 are presented, or only the ones having a similarity 803 of “0.7 or more” are presented. ..

From the presented modification candidates, the user or the application that called the search system identifies the modification candidate by selecting an appropriate one (corresponding to the identification means). For example, suppose "big data" is selected.

In step S306, the correction result receiving unit 108 receives the selection result of the user or the application, replaces the word with the spelling error in the input character string with the selected result, and corrects it (corresponds to the correction means). .. Specifically, "Big Day?" Is replaced with "Big Data" to correct it.

In step S307, the document search unit 102 searches the document storage unit 121 again with the replaced search conditions (corresponding to the search means). Specifically, the search is performed again under the three conditions of "artificial intelligence", "machine learning", and "big data". As a result, the text document 501b is ranked lower, and the results of 501c and 501d are ranked higher.

Further, although not shown, the text document 501 containing only "big data" out of the three conditions is not hit in the search in step S302, but is hit in the re-search in step S307.

In step S308, a list of text documents 501 whose ranks have been changed or which have not been hit in the first time is presented as a result of step S307 which is re-searched based on the search condition modified in step S306. ..

As described above, in steps S305 and S306, it has been explained that the correct spelling word is selected by the user or the application from the correction candidates for the word having the spelling error. However, this is just an example. In the search system of the present invention, the one with the highest degree of similarity may be selected by default.

For example, if a re-search is performed without presenting the result of the first search or the result of the spelling check to the user and the result of the re-search is presented up to step S308, the user spells in the input character string. It makes it look like you've searched with the correct spelling, even though you've made a mistake. This completes the description of the flowchart of FIG.

It should be noted that the structure and contents of the various data described above are not limited to this, and it goes without saying that the structure and contents are various depending on the intended use and purpose.

According to the present invention, it is possible to search the document data by appropriately modifying the modified character string included in the character string received from the user.

Although some embodiments have been described above, the present invention can take an embodiment as, for example, a system, an apparatus, a method, a computer program, a recording medium, or the like, and specifically, a plurality of devices. It may be applied to a system composed of, or may be applied to a device composed of one device.

Further, the computer program in the present invention is a computer program in which a computer can execute the processing methods of the flowcharts shown in FIGS. 3 to 4, and the storage medium of the present invention can execute the processing methods in FIGS. 3 to 4. Computer programs are stored. The computer program in the present invention may be a computer program for each processing method of each device of FIGS. 3 to 4.

As described above, a computer in which a recording medium on which a computer program that realizes the functions of the above-described embodiment is recorded is supplied to the system or device, and the computer (or CPU or MPU) of the system or device is stored in the recording medium. Needless to say, the object of the present invention can be achieved by reading and executing the program.

In this case, the computer program itself read from the recording medium realizes the novel function of the present invention, and the recording medium storing the computer program constitutes the present invention.

Recording media for supplying computer programs include, for example, flexible disks, hard disks, optical disks, magneto-optical disks, CD-ROMs, CD-Rs, DVD-ROMs, magnetic tapes, non-volatile memory cards, ROMs, EEPROMs, etc. Silicon disks, solid state drives, etc. can be used.

Further, by executing the computer program read by the computer, not only the functions of the above-described embodiment are realized, but also the OS (operating system) or the like running on the computer is activated based on the instructions of the computer program. Needless to say, there are cases where a part or all of the actual processing is performed and the processing realizes the functions of the above-described embodiment.

Further, the computer program read from the recording medium is written to the memory provided in the function expansion board inserted in the computer or the function expansion unit connected to the computer, and then its function is based on the instruction of the computer program code. Needless to say, there are cases where a CPU or the like provided in the expansion board or the function expansion unit performs a part or all of the actual processing, and the functions of the above-described embodiment are realized by the processing.

Further, the present invention may be applied to a system composed of a plurality of devices or a device composed of one device. It goes without saying that the present invention can also be applied when it is achieved by supplying a computer program to a system or device. In this case, by reading a recording medium containing a computer program for achieving the present invention into the system or device, the system or device can enjoy the effects of the present invention.

Further, by downloading and reading a computer program for achieving the present invention from a server, database, or the like on the network by a communication program, the system or device can enjoy the effect of the present invention.

It should be noted that all the configurations in which each of the above-described embodiments and modifications thereof are combined are also included in the present invention.

100 Information processing device 101 Input character string reception unit 102 Document search unit 103 Word extraction unit 104 Dictionary search unit 105 Similarity judgment unit 106 Correction candidate storage unit 107 Correction candidate presentation unit 108 Correction result reception unit 121 Document storage unit 122 Word dictionary storage Part 311 Extracted word storage unit 312 Correction candidate storage unit 600 Co-occurrence information storage unit

Claims (8)

An information processing device that acquires document data searched for search conditions including character strings from a plurality of document data.
In the above search conditions, a specific means for identifying an erroneous character string and
The acquired document data includes a character string that corrects the erroneous character string based on the character string included in the document data acquired for the search condition and the specified erroneous character string. and a selecting means for selecting from the character string,
The selection means selects the character string to be corrected based on the value related to the coexistence of the character string included in the document data and the character string included in the search condition in the acquired document data. An information processing device that features it. The information processing apparatus according to claim 1, further comprising an output means for outputting the selected character string so as to display the selected character string. The information processing apparatus according to claim 1 or 2, further comprising a receiving means for receiving a designation of a character string for correcting the erroneous character string from the selected character string. The information processing apparatus according to any one of claims 1 to 3, wherein the document data re-searched for the search condition including the character string corrected by the character string to be corrected is acquired. The selection means is characterized in that the character string to be corrected is selected based on the degree of similarity between the character string included in the document data acquired for the search condition and the character string having the specified error. The information processing apparatus according to any one of claims 1 to 4. The value related to the co-occurrence is a number based on the frequency of co-occurrence between the character string included in the document data and the character string included in the search condition in the acquired document data, and is included in the document data. The information processing apparatus according to any one of claims 1 to 5, wherein the weighting coefficient is related to the selection of the character string to be used. It is a control method of an information processing device that acquires document data searched for a search condition including a character string from a plurality of document data.
A specific step in which the specific means identifies an erroneous character string in the search condition, and
The selection means has acquired the character string for correcting the erroneous character string based on the character string included in the document data acquired for the search condition and the specified erroneous character string. It has a selection step to select from the character strings included in the document data .
In the selection step, the character string to be corrected is selected based on the value related to the coexistence of the character string included in the document data and the character string included in the search condition in the acquired document data. A characteristic method of controlling an information processing device. A program that can be executed by an information processing device that acquires document data searched for search conditions including character strings from multiple document data.
The information processing device
In the above search conditions, a specific means for identifying an erroneous character string and
The acquired document data includes a character string that corrects the erroneous character string based on the character string included in the document data acquired for the search condition and the specified erroneous character string. to function as a selecting means for selecting from the character string,
The selection means selects the character string to be corrected based on the value related to the coexistence of the character string included in the document data and the character string included in the search condition in the acquired document data. Characterized program.

Images