JP2011141681A - Similar word search server and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a similar word search server and method for searching a similar word without omission by more increasing the width of the enlargement of a search range. <P>SOLUTION: A similar word search server 1 includes: a reading-down search means 12 for performing the search of a candidate word based on the comparison of key input information received by using a dictionary DB21 in response to the reception of the key input information of a character string designated as a search query; a first result output means 13 for outputting a plurality of first candidate words as the result of performance by the reading-down search means 12; a character base search means 14 for performing, with respect to each of the plurality of first candidate words output by the first result output means 13, the search of the candidate words based on the comparison of the character information of characters corresponding to the first candidate words by using the dictionary DB21; and a second result output means 15 for outputting a plurality of second candidate words as the performance result by the character base search means 14. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a similar word search server and method.

  2. Description of the Related Art Conventionally, a document search apparatus that searches a document database that matches a search condition has been disclosed (for example, Patent Document 1). The document search apparatus described in Patent Literature 1 searches not only words included in a query but also synonyms and related words. As a method, the document search device extracts a word by morphologically analyzing a query, identifies a set of words having high co-occurrence by collating with co-occurrence information in the co-occurrence information file, and based on the word expansion result information Expand to include synonyms or related terms.

JP 2008-203997 A

  The expanding synonyms and related words described in Patent Document 1 are based on past processing objects, and therefore lack information on words with low popularity, and the range of search range expansion is limited.

  An object of the present invention is to provide a similar word search server and method for searching for similar words without omission by increasing the range of expansion of the search range.

  The inventors of the present invention have found that the search range as candidates is widened by reading down the search query and that the search is performed for each of the candidates, and the present invention has been completed. Specifically, the present invention provides the following.

(1) A read-out search means for executing a search for a candidate word based on a comparison of the key input information received using the dictionary DB in response to receiving key input information of a character string designated as a search query;
First result output means for outputting a plurality of first candidate words as an execution result by the reading-down search means;
Retrieval of candidate words based on comparison of character information of characters corresponding to the first candidate word using the dictionary DB for each of the plurality of first candidate words output by the first result output means A character-based search means for executing
Second result output means for outputting a plurality of second candidate words as execution results by the character-based search means;
Comprising
Similar word search server.

  According to such a configuration of the present invention, a candidate word search is performed based on a comparison of key input information of a character string specified as a search query, and a plurality of first candidate words output as the execution result are searched. For each of these, a candidate word search is performed based on a comparison of character information of characters corresponding to the first candidate word. Therefore, for example, when the Kana-Kanji conversion error or Kanji character is unknown and a reading is input, the similar word search server covers possible reading patterns and further searches including a character-based similar word. Even when there is a mistake in kanji conversion, similar word candidates can be extracted with less omission.

(2) The read-down search means calculates a similarity obtained by quantifying the search accuracy of the candidate word,
The first result output means outputs the plurality of first candidate words whose similarity satisfies a predetermined criterion.
The similar word search server according to (1).

  According to such a configuration of the present invention, when a similarity is obtained by quantifying the accuracy of search for candidate words, and the similarity satisfies a predetermined criterion, a plurality of first candidates that satisfy the criterion Output words. Thus, the similar word search server can search for candidate words while adjusting the expansion range of the search range.

(3) The second result output means outputs the plurality of second candidate words by the character-based search means in descending order of the degree of similarity calculated by the read-down search means.
The similar word search server according to (2).

  According to such a configuration of the present invention, the second candidate word is output by the character-based search means in order from the highest degree of similarity. Therefore, the similar word search server can output candidate words in descending order of accuracy according to the similarity.

(4) receiving key input information of a character string designated as a search query;
Performing a search for candidate words based on a comparison of the key input information received using a dictionary DB;
Outputting a plurality of first candidate words as an execution result by searching for candidate words based on the comparison of the key input information;
Performing a search for candidate words based on a comparison of character information of characters corresponding to the first candidate words using the dictionary DB for each of the plurality of output first candidate words;
Outputting a plurality of second candidate words as an execution result by searching for candidate words based on the comparison of the character information;
including,
Similar word search method.

  According to the present invention, a candidate word search is performed based on a comparison of key input information of a character string specified as a search query, and character information of characters corresponding to a plurality of candidate words output as the execution result is compared. Since the search for candidate words based on is performed, it is possible to extract similar word candidates with a wider search range and less leakage.

It is a figure which shows the whole structure of the similar word search system which concerns on this embodiment, and the functional structure of a similar word search server. It is a flowchart of the main process of the similar word search server which concerns on this embodiment. It is a figure which shows the variation of the key input information which concerns on this embodiment. It is a figure which shows the specific example of the search of the candidate word which concerns on this embodiment.

  DESCRIPTION OF EMBODIMENTS Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. This is merely an example, and the technical scope of the present invention is not limited to this.

(Embodiment)
[Overall Configuration of Similar Word Search System 100 and Functional Configuration of Similar Word Search Server 1]
FIG. 1 is a diagram showing the overall configuration of the similar word search system 100 and the functional configuration of the similar word search server 1 according to the present embodiment.

  The similar word search system 100 includes a similar word search server 1, a user terminal 3, and a communication network 9.

  The similar word search server 1 is a server that searches for similar words and transmits the candidate words to the user terminal 3 in response to a search request for candidate words from the user terminal 3. The similar word search server 1 includes a control unit 10 and a storage unit 20.

  The control unit 10 includes a search request receiving unit 11, a reading-out search unit 12, a first result output unit 13, a character base search unit 14, and a second result output unit 15.

  The search request receiving unit 11 is a control unit that receives key input information of a character string designated as a search query transmitted from the user terminal 3. The key input information refers to key information itself that is input when the user performs keyboard input on the user terminal 3. In the user terminal 3, when the user inputs Roman characters in the search window displayed on the Web browser, for example, when "yosano" is entered as "yosano" and kanji conversion is made into "yosano", the search request receiving means 11 Receives the input of “yosano”.

  The reading-down search unit 12 is a control unit that searches for similar words using a dictionary DB 21 (DB: database) based on readings. Specifically, the reading-down search unit 12 searches for similar words based on the comparison of the received key input information. When “yosano” is input as “yosano” at the user terminal 3, the read-down search unit 12 searches for similar words from “yosano” received from the user terminal 3. Since the reading method is appropriate, the reading-down search means 12 can search appropriately. On the other hand, if the user inputs “yosano” to “yosano” at the user terminal 3, the reading is incorrect, and the search means 12 reads the “yosano” appropriately from “yosano” received from the user terminal 3. ”Cannot be searched.

  The first result output means 13 is a control unit that executes a search for candidate words based on a comparison of key input information. The first result output means 13 is a control unit that reads out and outputs a search request candidate word (first candidate word). The first result output means 13 outputs n results (n is an integer) from result 1 to result n as candidate words. Since the candidate word is a result of a reading-down search by reading and is a search by romaji input, there is almost no word that does not match the pattern in the romaji character string, and there is almost no zero match.

  Further, the first result output means 13 calculates a similarity obtained by quantifying the accuracy of the search for the candidate word based on each of the plurality of input source candidate words. For example, the first result output unit 13 outputs candidate words whose calculated similarity satisfies a predetermined criterion (for example, 0.5 or more, where 1 is conforming and 0 is not conforming).

  The character-based search means 14 is a control unit that searches for candidate words based on a comparison of character information of characters corresponding to candidate words such as kanji and hiragana. The character-based search unit 14 uses the dictionary DB 21 for each of the results 1 to n, which are a plurality of candidate words output as candidate words by the first result output unit 13, and candidates based on comparison of character information Search for a word.

  The second result output means 15 is a control unit that outputs a candidate word (second candidate word) by character-based search. The second result output means 15 obtains the result 1a (a is an integer) from the result 11 as a candidate word with the result 1 as an input,... Output multiple results up to (integer). The second result output unit 15 outputs the candidate words so that the similarity calculated by the first result output unit 13 is in descending order.

  The storage unit 20 includes a dictionary DB 21. The dictionary DB 21 is a DB that stores various dictionaries including a thesaurus that is a dictionary of similar words. Other various dictionaries include, for example, a national language dictionary, a new word dictionary, an English-Japanese dictionary, a Japanese-English dictionary, and an encyclopedia.

  The hardware of the similar word search server 1 of this embodiment may be configured by a general computer. For example, a general computer includes a central processing unit (CPU) as the control unit 10 and a memory (RAM, ROM), a hard disk (HDD), and an optical disk (CD, DVD, etc.) as a storage unit 20 in a network. As a communication device, various wired and wireless LAN devices, as a display device, for example, various displays such as a liquid crystal display and a plasma display, and as an input device, for example, a keyboard and a pointing device (mouse, tracking ball, etc.) are appropriately provided. These are connected by a bus line. In such a general computer, the CPU controls the similar word search server 1 in an integrated manner, reads and executes various programs as appropriate, and cooperates with the hardware described above, thereby providing various functions according to the present invention. Is realized.

  The user terminal 3 is, for example, a personal computer (PC) or a mobile terminal such as a mobile phone. The user terminal 3 may be any device as long as it has a communication function and can input characters and display a web page.

  The communication network 9 is a communication line such as the Internet for performing communication between the similar word search server 1 and the user terminal 3. The communication network 9 may be wired or part or all of it may be wireless.

[Process of Similar Word Search Server 1]
Next, processing in the similar word search server 1 will be described. FIG. 2 is a flowchart of the main process of the similar word search server 1 according to the present embodiment. FIG. 3 is a diagram showing variations of key input information according to the present embodiment. FIG. 4 is a diagram showing a specific example of candidate word search according to the present embodiment.

  S <b> 1 in FIG. 2: the control unit 10 (search request receiving means 11) receives a search request from the user terminal 3. The search request includes key input information of a character string specified as a search query.

  Here, variations of key input information are shown. FIG. 3 shows a variation of key input when trying to input “Naoe Kanetsugu”, FIG. 3 (1) shows a variation of Roman input, and FIG. 3 (2) shows a variation of Kana input. FIG. 3A shows two patterns of correct input and one incorrect input as an example. “Tsu” in Japanese may be either “tsu” or “tu” for Roman input. The example of incorrect input represents a case where the keyboard input order is incorrect. FIG. 3B shows one correct input and one incorrect input as an example. In the example of an incorrect input, the case where “ka” is input as “ha” by mistake is shown.

  For example, when the user erroneously inputs “naokeanetugu”, the control unit 10 determines that the key input information may be mistaken for “naoekanetugu” or “naoekanetsugu”. , “Naoekanetugu” is also prepared for processing.

  Returning to FIG. 2, the control unit 10 (read-down search unit 12) uses the dictionary DB 21 to search for candidate words based on the comparison of key input information. Searching for candidate words based on a comparison of key input information refers to candidate words based on variations in the keyboard alphabet. FIG. 4 shows a case where the user inputs “Naoe Kanetsugu” as a search query. In this case, as illustrated in FIG. 4A, the control unit 10 performs a search by reading down using “naoekanetugu” input in Roman characters. As a read-down search, the control unit 10 may output a candidate word “Naoekanetsugu” that is, for example, just converted from a Roman character input into kana.

  Returning to FIG. 2, S3: The control unit 10 (read-down search means 12) determines whether or not the search result in S2 includes hiragana characters or katakana. When the hiragana character or the katakana character is included (S3: YES), the control unit 10 moves the process to S4. On the other hand, when the hiragana character or the katakana character is not included (S3: NO), the control unit 10 moves the process to S5.

  S4: The control unit 10 converts hiragana characters or katakana characters into kanji.

  S5: The control unit 10 (first result output means 13) outputs a search result of candidate words based on the comparison of the key input information. Here, the control unit 10 outputs both the hiragana or katakana characters before conversion into kanji and the characters after conversion into kanji as search results. In addition, the control unit 10 outputs, as search results, the top five items having a degree of similarity of 0.5 or more as candidate words. FIG. 4B shows a character string that appears as a candidate word when read down, and a hiragana character first, a kanji converted character string, and then a descending score. Thus, by ranking according to the degree of similarity, the similar word search server 1 can start the subsequent processing in order from candidate words accurate to the key input information.

  Returning to FIG. 2, S <b> 6: The control unit 10 (character-based search unit 14) is based on comparison of character information of characters corresponding to each of candidate words output using the dictionary DB 21 or comparison of input character information. To search for candidate words. FIG. 4C shows a result of performing a character base search using character information of “Kaneki Naoe”, “Kanego Konoe”,.

  Returning to FIG. 2, S7: The control unit 10 (second result output unit 15) determines the candidate word search results in descending order of the degree of similarity of the candidate word results based on the comparison of the key input information. A result page that outputs candidate words is generated and transmitted to the user terminal 3. In the example shown in FIG. 4, first, candidate words related to “Naoe Kanetsugu” having a similarity of “1” are output, and then candidate words related to “Naoe Kanetsu” having a similarity of “1”. Output. Then, the control part 10 complete | finishes this process.

  In the example of FIG. 4, based on the search query “Naoe Kanetsugu” having key input information of “naoekanetugu”, the read-out search means 12 outputs “Konoe Kanetsugu” having key input information of “naoekanetugu”, Furthermore, the character-based search means 14 outputs “Konbe Bungei”. As described above, the similar word search server 1 can be obtained when a plurality of candidate words that have been read down are input and a character-based search is further performed to perform a read-down search or a character-based search alone. Can output difficult “Konoe Bungei”.

  As described above, the similar word search server 1 executes a read-out search based on a comparison of key input information of a Roman character input of a character string specified as a search query, and a plurality of candidate words output as the execution result. This time, the candidate word search is executed based on the comparison of the character information of the characters corresponding to the candidate word. Therefore, the similar word search server 1 can search for each of the candidates by increasing the width of the search range as candidates based on the search query.

  And, for example, when a kana-kanji conversion error or kanji is unknown and a reading is input, the similar word search server 1 covers a possible reading pattern and further searches including a character-based similar word. Even when there is a kanji conversion error, it is possible to extract similar word candidates with few omissions.

(Deformation)
In the present embodiment, a personal computer that mainly performs keyboard input has been described as an example. However, a mobile phone that changes words by repeatedly pressing buttons may be used.

  As mentioned above, although embodiment of this invention was described, this invention is not restricted to embodiment mentioned above. The effects described in the embodiments of the present invention are only the most preferable effects resulting from the present invention, and the effects of the present invention are limited to those described in the embodiments of the present invention. is not.

DESCRIPTION OF SYMBOLS 1 Similar word search server 3 User terminal 10 Control part 11 Search request receiving means 12 Reading down search means 13 1st result output means 14 Character base search means 15 2nd result output means 20 Storage part 21 Dictionary DB
100 Similar word search system

Claims (4)

A read-out search means for performing a search for candidate words based on a comparison of the key input information received using the dictionary DB in response to receiving key input information of a character string designated as a search query;
First result output means for outputting a plurality of first candidate words as an execution result by the reading-down search means;
Retrieval of candidate words based on comparison of character information of characters corresponding to the first candidate word using the dictionary DB for each of the plurality of first candidate words output by the first result output means A character-based search means for executing
Second result output means for outputting a plurality of second candidate words as execution results by the character-based search means;
Comprising
Similar word search server. The read-down search means calculates a similarity obtained by quantifying the search accuracy of the candidate word,
The first result output means outputs the plurality of first candidate words whose similarity satisfies a predetermined criterion.
The similar word search server according to claim 1. The second result output means outputs the plurality of second candidate words by the character base search means in descending order of the degree of similarity calculated by the read-down search means.
The similar word search server according to claim 2. Receiving key input information of a character string specified as a search query;
Performing a search for candidate words based on a comparison of the key input information received using a dictionary DB;
Outputting a plurality of first candidate words as an execution result by searching for candidate words based on the comparison of the key input information;
Performing a search for candidate words based on a comparison of character information of characters corresponding to the first candidate words using the dictionary DB for each of the plurality of output first candidate words;
Outputting a plurality of second candidate words as an execution result by searching for candidate words based on the comparison of the character information;
including,
Similar word search method.

Images