JP2010134922A - Similar word determination method and system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a similar word determination method and system. <P>SOLUTION: The similar word determination method includes steps of: determining whether an input character string is a first language or a second language; calculating editing distance between a character string by writing pronunciation of a candidate character string in the second language among candidate character strings written in the first language and the input character string when the input character string is the second language; and determining the candidate character string written in the first language corresponding to the character string written in the second language whose editing distance with the input character string is equal to or less than a reference value among the candidate character strings as a similar word of the input character string. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a search service, and more particularly, to a method and system for providing a similar word as a recommended query to a foreign language or a Korean expression of a corresponding foreign language pronunciation.

  Recently, with the development of science and technology and the improvement of economic standards, the spread of communication networks such as the ultra-high-speed Internet and the number of users of ultra-high-speed communication networks are increasing rapidly. The increase allowed the development of new services and the diversification of service items via the communication network. It can be said that the most common service among services using such a communication network is a search service.

  When a query is entered by a user, a search service is a search result corresponding to the entered query (for example, a website containing the entered query, an article containing the entered query, or an entered query). A service that provides a user with an image having a file name to include.

  However, when using a search service, a user cannot enter a query correctly because he / she enters the query incorrectly, or he / she does not know exactly what query he / she wants. In some cases, it is not possible to accurately input a query desired to be input. In such a case, the search service provider performs a search based on the actually inputted query, and as a result, there is an inconvenience that the user cannot receive the search result desired by the user.

  In order to solve such inconvenience, recent search services provide various search services such as providing a recommended query for a query input by a user or providing a related query. Here, the provision of a recommended query means a service that provides a part of a query similar to a query input by a user as a recommended query.

  In particular, in providing such a recommended query, if the user is not proficient in a foreign language and does not know the correct foreign language notation or pronunciation, the foreign language notation of the query that the user is trying to search for Alternatively, there is a problem in that the user cannot correctly provide a desired search result by erroneously inputting the corresponding Korean pronunciation of the foreign language pronunciation.

  The present invention is for solving the above-described problem, and even when a foreign language expression or pronunciation corresponding to a query to be searched by a user is not accurately understood, a similar foreign language character string or similar It is an object of the present invention to provide a similar word determination method and system that can provide a foreign language character string corresponding to a pronunciation as a recommended query.

  In addition, the present invention provides a foreign language whose pronunciation is similar even when the query to be searched by a user is a foreign language, even when the corresponding foreign language is unknown but the native language notation of the corresponding foreign language pronunciation is known. It is another technical problem to provide a method and system for determining a similar word that can be provided as a recommended query.

  Furthermore, the present invention provides a similar word determination method and system capable of providing similar words as recommended queries even when a query to be searched by a user is in various languages such as Chinese and English. This is another technical issue.

  The similar word determination method according to an embodiment of the present invention includes a step of determining whether an input character string is a first language or a second language, and when the input character string is the second language, Calculating an edit distance between a character string in which the pronunciation of the candidate character string is written in the second language and the input character string among candidate character strings written in the first language; and The candidate character string written in the first language corresponding to the character string written in the second language whose edit distance from the input character string is not more than a reference value is determined as a similar word of the input character string Including the step of.

  Here, when the second language is a native language, the character string in which the pronunciation of the candidate character string is expressed in the second language is converted by the sound difference converter for the candidate character string with respect to the native language. Preferably, when the second language is Korean, the sound difference converter for the native language is a Korean sound difference converter. Here, the sound difference converter can convert the first language into the phonetic notation of the native language that is the second language.

  On the other hand, in the determination step, when the input character string is in the first language, a step of calculating an edit distance between the candidate character string and the input character string, and the edit distance in the candidate character string is Determining a candidate character string that is equal to or less than a reference value as a similar word of the input character string.

  The step of calculating the edit distance further includes a step of determining whether or not the input character string made of the second language can be converted to the first language, and the input character made of the second language When a string cannot be converted into the first language, an edit distance between a character string representing the pronunciation of the candidate character string in the second language and the input character string is calculated.

  Then, in the similar word determination method according to an embodiment of the present invention, when the input character string composed of the second language can be converted into the first language, the input character string composed of the second language is converted into the first character string. A step of converting into one language, a step of calculating an edit distance between the character string converted into the first language and the candidate character string, and a candidate whose edit distance is not more than a reference value in the candidate character string And a step of determining a character string as a similar word of the input character string.

  At this time, in the step of converting, when the first language is a foreign language, the input character string is converted into a character string composed of the corresponding foreign language using a sound difference restoration device for the corresponding foreign language. It is characterized by. In one embodiment, when the first language is Japanese, the sound difference restorer for the corresponding foreign language can be a Japanese sound difference restorer.

  In one embodiment, the first language may be any one of foreign languages, and the second language may be a native language. Further, the input character string and the candidate character string can be a search query.

  Meanwhile, the similar word determination method according to an embodiment of the present invention further includes a step of receiving the input character string from a user terminal before the determination step, and the determined similarity is after the similar word determination step. The method further includes providing a word as a recommended query to the user terminal.

  In one embodiment, the candidate character string includes a candidate character string having an edit distance with respect to the input character string that is equal to or less than a reference value among candidate character strings stored in advance or a candidate that includes characters common to the input character string The character string is selected from at least one candidate character string having a character similarity score within the upper N rank among the character strings.

  Here, a candidate character string whose edit distance is equal to or less than a reference value among the candidate character strings is selected using a wild card character search for each operation for calculating the edit distance. The candidate character string including characters common to the input character string is a candidate character string including N-gram common to the input character string, and the character similarity score is common to the input character string N-gram size, the number of the common N-grams, the similarity of the position where the common N-gram is found, and the length difference between the input character string and each candidate character string It is determined by using.

  To achieve the above object, a similar word determination system according to an embodiment of the present invention receives an input character string from a user terminal and provides the user terminal with a similar word corresponding to the input character string as a recommended query. When the user interface unit and the input character string are in the second language, the character string expressing the pronunciation of the candidate character string in the second language and the input character among the candidate character strings expressed in the first language A candidate character string written in the first language corresponding to a character string written in the second language whose edit distance to the column is equal to or less than a reference value, and determining the candidate character string written in the first language as a similar word of the input character string; And a similar word determination unit provided to

  According to the present invention, when a query to be searched by a user is a foreign language, even when an accurate foreign language is not known, a foreign language having a similar notation can be provided as a recommended query. .

  In addition, the present invention provides a foreign language whose pronunciation is similar even when the query to be searched by a user is a foreign language, even when the corresponding foreign language is unknown but the native language notation of the corresponding foreign language pronunciation is known. There is an effect that it can be provided as a recommended query.

  Furthermore, the present invention has an effect that similar words can be provided as a recommended query even when a query to be searched by a user is in various languages such as Chinese and English.

1 is a block diagram schematically showing a similar word determination system according to an embodiment of the present invention. It is a figure which shows the detailed structure of the similar word determination part shown in FIG. It is a flowchart which shows the similar word determination method which concerns on one Embodiment of this invention.

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

  FIG. 1 is a diagram illustrating a network configuration including a similar word determination system according to an embodiment of the present invention. As shown in the figure, the similar word determination system 100 receives an input character string from a user terminal 120 connected via the Internet 110, determines a similar word for the received input character string, and recommends the determined similar word to the recommended query. To the user terminal 120. Such a similar word determination system 100 includes a user interface unit 130, a candidate character string providing unit 140, and a similar word determination unit 150, as shown in the figure.

  First, the user interface unit 130 receives an input character string written in a foreign language or an input character string in which the pronunciation of the foreign language is written in Korean from the user terminal 120, and receives a similar word from the similar word determination unit 150 described later. The word is provided to the user terminal 120 as a recommended query for the input character string.

  The candidate character string providing unit 140 provides the candidate character string to the similar word determining unit 150 in order to calculate the edit distance with the input character string for determining the recommended query to be provided to the user. The candidate character strings in the present invention are those in which some of the pre-stored queries are selected as candidate character strings, and the similar word determination unit 150 inputs all the pre-stored queries as targets. Rather than calculating the character string and edit distance, the response speed of the similar word providing service to the input query is improved by calculating the edit distance between the candidate character string provided from the candidate character string providing unit 140 and the input character string. can do.

  Here, the candidate character string provided by the candidate character string providing unit 140 is input in a candidate character string whose edit distance to the input character string is equal to or less than a reference value or a candidate character string including characters common to the input character string At least one of the candidate character strings whose character similarity score with the character string is within the upper N ranks can be stored in advance in a database (not shown).

  A candidate character string whose edit distance is equal to or less than a reference value among the candidate character strings can be selected using a wild card character search for each operation for calculating the edit distance.

  Here, each operation includes an insertion operation, a deletion operation, an exchange operation, and a transposition operation, and the insertion operation means an operation that occurs when a new character is added to a specific character string. An operation means an operation generated by deleting a character included in a specific character string, and an exchange operation means an operation generated by exchanging a character included in a specific character string with a new character. The transposition operation means an operation generated by changing the order of adjacent characters included in a specific character string.

  A candidate character string including a character common to the input character string is a candidate character string including an N-gram common to the input character string, and the character similarity score is a large N-gram common to the input character string. It can be determined using the number of common N-grams, the similarity of the position where the common N-gram is found, and the difference in length between the input character string and each candidate character string.

  When the input character string is a character string written in the second language corresponding to the pronunciation of the first language, the similar word determination unit 150 includes the candidate character strings stored in advance provided by the candidate character string providing unit 140. The candidate character string corresponding to the character string expressed in the second language whose edit distance between the character string expressed in the second language corresponding to the pronunciation of the candidate character string and the input character string is equal to or less than the reference value is input character string And are provided to the user interface unit 130. For this purpose, as shown in FIG. 2, the similar word determination unit 150 includes an input character string determination unit 210, a character string conversion necessity determination unit 220, a first editing distance calculation unit 230, and a first determination unit 240. A second edit distance calculation unit 250, a second determination unit 260, and a character string conversion unit 270. Hereinafter, the similar word determination unit 150 will be described in detail with reference to FIG.

  The input character string determination unit 210 determines whether the input character string is a character string written in the first language or a character string written in the second language corresponding to the pronunciation of the first language. In one embodiment, the first language can be any one of the foreign languages and the second language can be the native language. For example, in the Republic of Korea, the first language can be one of a variety of foreign languages such as Japanese, Chinese, and English, and the second language should be the native language, Korean. Can do. If the first language is Japanese, the input character string may include at least one of hiragana, katakana, and kanji. Here, the input character string and the candidate character string can be a search query.

  In one embodiment, the input character string determination unit 210 is a character string written in the first language or a character string written in the second language, depending on the character code of each character of the input character string. It can be judged whether there is. For example, assuming that the first language is Japanese and the second language is Korean, the input character string determination unit 210 checks the character codes of all syllables in the input character string, and all characters are in Korean. It can be determined that the input character string is Korean only when the input character string is Japanese. If both Japanese and Korean are written, it can be determined that the input character string is Japanese. More preferably, each character of the input character string is converted into a UCS-2 code, and if the Unicode value exists in the area of 0xAC00 and 0xD7A3, the input character string can be determined as Korean.

  On the other hand, in the present invention, the first language and the second language are not limited to the languages described in the present invention, and the first language and the second language can be various languages. For convenience of explanation, the following description will be made assuming that the first language is Japanese and the second language is Korean.

  In other words, the input character string determination unit 210 determines whether the character string input via the user terminal is a Japanese character string or whether the Japanese pronunciation is a character string written in Korean.

（「タップリ」の韓国語表記、以下同じ）である場合には、該当する日本語表記である「たっぷり」が存在しているため日本語への変換が可能であると判断し、入力文字列が

（「タップルリ」の韓国語表記、以下同じ）である場合には、該当する日本語表記が存在しないため日本語への変換が不可能であると判断する。 When the character string conversion necessity determination unit 220 determines that the input character string is an input character string expressed in the second language by the input character string determination unit 210, the input character string is expressed in the first language. It is determined whether or not conversion to a character string is possible. For example, if the input string is

If it is (Korean notation for “Tappuri”, the same applies hereinafter), it is determined that conversion to Japanese is possible because the corresponding Japanese notation “Plenty” exists, and the input character string But

If it is (Korean notation for “Tapluri”, the same applies hereinafter), it is determined that conversion to Japanese is impossible because there is no corresponding Japanese notation.

である場合、入力文字列である

と候補文字列の発音である韓国語文字列との編集距離を算出する。 When the input character string determination unit 210 determines that the input character string is a character string described in the second language, the first edit distance calculation unit 230 displays the second edit distance in the second language corresponding to the pronunciation of the candidate character string. The edit distance between the inputted character string and the input character string is calculated. For example, the input string is a string whose pronunciation is written in Korean

Is the input string

And the edit distance between the Korean character string that is the pronunciation of the candidate character string.

のような韓国語で表記された入力文字列が日本語に変換が不可能である場合には、

と候補文字列の発音に該当する

、

（「タッスル」の韓国語表記、以下同じ）、

（「タタエル」の韓国語表記、以下同じ）などのような文字列と編集距離を算出する。 In one embodiment, the first edit distance calculation unit 230 determines that the character string conversion necessity determination unit 220 cannot convert an input character string written in the second language into a character string written in the first language. In this case, the edit distance between the character string written in the second language corresponding to the pronunciation of the candidate character string and the input character string can be calculated. In other words, as mentioned above,

If the input string written in Korean cannot be converted to Japanese,

Corresponds to the pronunciation of the candidate string

,

(Korean notation for "Tassle", the same applies below),

(Tatael is written in Korean, the same applies hereinafter) and the edit distance are calculated.

の場合、

と編集距離が基準値以下である

、

、

に該当する候補文字列である「たっぷり」、「たっする」、「たたえる」を入力文字列の類似語として決定することができる。一方、本発明における編集距離の基準値は、状況によって変更が可能であり、特定の数値に限定されるものではない。 The first determination unit 240 determines a candidate character string corresponding to a character string written in a second language whose edit distance to the input character string is equal to or less than a reference value among the candidate character strings as a similar word of the input character string To do. For example, the input string is a string written in Korean

in the case of,

And edit distance is below the reference value

,

,

Candidate character strings corresponding to ‘full’, ‘tact’, and ‘come’ can be determined as similar words in the input character string. On the other hand, the edit distance reference value in the present invention can be changed depending on the situation, and is not limited to a specific numerical value.

に変換することができる。 Here, when the second language is Korean, the character string written in the second language corresponding to the pronunciation of the candidate character string can be obtained using a Korean sound difference converter for the candidate character string. As an example, the sound difference converter can convert a candidate character string into a character string composed of phonetic notation of the second language. For example, if the candidate character string is Japanese and the second language is Korean, the sound difference converter can convert Japanese into Korean phonetic notation. Specifically, if the candidate character string in Japanese is “full”, the tone difference converter is a character string composed of “full” in Korean phonetic notation.

Can be converted to

  When the input character string determining unit 210 determines that the input character string is a character string written in the first language, the second edit distance calculating unit 250 calculates an edit distance between the candidate character string and the input character string. For example, if the input character string is input as “TAPUL” which is Japanese, the edit distance between the candidate character string and “TAPLUL” which is the input character string is calculated.

  The second determination unit 260 determines a candidate character string whose edit distance calculated by the second edit distance calculation unit 250 is equal to or less than a reference value among the candidate character strings as a similar word of the input character string. For example, if the input character string is Japanese “Tapuru”, “Candidate” is a candidate character string whose edit distance between the candidate character string and the input character string “Tapuru” is less than or equal to the reference value, “Tat” and “Respond” are determined as similar words to the input character string “Tapuru”.

  If the character string conversion necessity determination unit 220 determines that the input character string written in the second language can be converted into a character string written in the first language, the character string conversion unit 270 An input character string expressed in a language is converted into a character string expressed in a first language. Here, the second edit distance calculation unit 250 may calculate the edit distance between the input character string converted into the character string written in the first language converted by the character string conversion unit 270 and the candidate character string. it can.

  In one embodiment, the character string conversion unit 270 may be a Japanese sound difference restorer that converts an input character string into a Japanese character string when the first language is Japanese.

  On the other hand, in the above-described embodiment, it has been described that the candidate character string providing unit 140 is included in the similar word determination system 100. However, in a modified example of the embodiment, the candidate character string providing unit 140 may be included in another system and the database (FIG. (Not shown) can take the place of that role.

  Hereinafter, the similar word determination method according to the present invention will be described with reference to FIG. FIG. 3 is a flowchart illustrating a similar word determination method according to an embodiment of the present invention.

  As shown in the figure, the character string input by the user is received via the user terminal (S300).

  Next, it is determined whether the input character string is the first language or the second language (S310). Here, the first language can be one of Japanese, Chinese, and English, and the second language can be Korean. In other words, it is determined whether the input character string is in Japanese or the Korean notation of the corresponding Japanese pronunciation. When the first language is Japanese, the input character string may include at least one of hiragana, katakana, and kanji.

である場合には、該当する日本語表記である「たっぷり」が存在しているため日本語への変換が可能であると判断し、入力文字列が

である場合には、該当する日本語表記が存在しないため日本語への変換が不可能であると判断することができる。 Next, when the input character string is in the second language, it is determined whether or not the input character string in the second language can be converted into the first language (S320). For example, if the input string is

, It is determined that conversion to Japanese is possible because the corresponding Japanese notation “plentiful” exists, and the input string is

In the case of, it can be determined that conversion to Japanese is impossible because there is no corresponding Japanese notation.

である場合、入力文字列である

と候補文字列の発音を韓国語で表記した文字列との編集距離を算出する。 Next, when the input character string cannot be converted into the first language, the character string expressing the pronunciation of the candidate character string in the second language and the input character string among the candidate character strings expressed in the first language The edit distance is calculated (S330). Here, the input character string and the candidate character string can be a search query. For example, the input string is a string whose pronunciation is written in Korean

Is the input string

And the edit distance between the candidate character string and the character string describing the pronunciation of the candidate character string in Korean.

  Here, the candidate character string is a candidate character string that includes a character that is common to the input character string or a candidate character string whose edit distance to the input character string is equal to or less than a reference value among the candidate character strings stored in advance. It can be set as at least one of the candidate character strings whose character similarity score with the input character string is within the top N.

  In addition, a candidate character string whose edit distance is equal to or less than a reference value among the candidate character strings can be selected using a wild card character search for each calculation for calculating the edit distance.

  A candidate character string including a character common to the input character string is a candidate character string including an N-gram common to the input character string, and the character similarity score is a large N-gram common to the input character string. The number of common N-grams, the similarity of the position where the common N-gram is found, and the difference in length between the input character string and each candidate character string can be determined. .

である場合、

と編集距離が基準値以下である

、

、

に該当する候補文字列である「たっぷり」、「たっする」、「たたえる」を入力文字列の類似語として決定することができる。 Next, the candidate character string expressed in the first language corresponding to the character string expressed in the second language whose edit distance to the input character string is equal to or less than the reference value among the candidate character strings is similar to the input character string (S340). For example, the input string is a string written in Korean

If it is,

And edit distance is below the reference value

,

,

Candidate character strings corresponding to ‘full’, ‘tact’, and ‘come’ can be determined as similar words in the input character string.

  Finally, the determined similar word is provided to the user terminal as a recommended query (S350).

  On the other hand, if the input character string is in the first language in step S310, the edit distance between the candidate character string and the input character string is calculated (S360), and the candidate whose edit distance is equal to or less than the reference value among the candidate character strings. The character string is determined as a similar word of the input character string (S370). For example, if the input character string is “Japanese” that is Japanese, “Tapuri”, which is a candidate character string whose edit distance between the candidate character string and the input character string “Taple” is below the reference value, “ “Tataru” and “Tameru” are determined as similar words to “Tapuru” which is the input character string.

  In step S320, if the input character string can be converted to the first language, the input character string made of the second language is converted to the first language (S380), and the input character string converted to the first language is converted. And the edit distance between the candidate character string and the candidate character string (S360).

  The similar word determination method described above may be recorded on a computer readable medium including program instructions for executing various operations realized by a computer. The recording medium may include program instructions, data files, data structures, etc. alone or in combination. The recording medium and the program instructions may be specially designed and configured for the purposes of the present invention, and may be known and usable by those skilled in the computer software art.

  Examples of the computer-readable recording medium include a hard disk, a floppy (registered trademark) disk, a magnetic medium such as a magnetic tape, an optical recording medium such as a CD-ROM and a DVD, and a magnetic medium such as a floppy disk. Included are optical devices and hardware devices specially configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. On the other hand, such a recording medium may be a transmission medium such as an optical or metal line or a waveguide including a carrier wave that transmits a signal for storing program instructions, data structures, and the like.

  Examples of the program instructions include not only machine language codes generated by a compiler but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware device described above may be configured to operate as one or more software modules to perform the operations of the above-described embodiments of the present invention, and vice versa.

  As described above, the present invention has been described with reference to some embodiments of the present invention, but the present invention is not limited thereto. Those skilled in the art can understand that the present invention can be variously modified and changed without departing from the spirit and scope of the present invention described in the claims. It will be possible.

100: Similar word determination system 110: Internet 120: User terminal 130: User interface unit 140: Candidate character string providing unit 150: Similar word determination unit

Claims (22)

Determining whether the input string is a first language or a second language;
When the input character string is the second language, among the candidate character strings expressed in the first language, the character string expressing the pronunciation of the candidate character string in the second language, and the input character string Calculating the edit distance between and
Among the candidate character strings, a candidate character string written in the first language corresponding to a character string written in the second language whose edit distance from the input character string is equal to or less than a reference value is input to the candidate character string. Determining as a string similarity word;
A similar word determination method comprising: When the second language is a native language,
The similar word determination method according to claim 1, wherein a character string in which the pronunciation of the candidate character string is expressed in the second language is converted by a sound difference converter for the native language. . In the determining step,
When the input character string is in the first language, calculating an edit distance between the candidate character string and the input character string;
Determining a candidate character string whose edit distance is equal to or less than a reference value among the candidate character strings as a similar word of the input character string;
The similar word determination method according to claim 1, further comprising: In the step of calculating the edit distance,
Determining whether or not the input character string in the second language can be converted into the first language;
When the input character string in the second language cannot be converted into the first language, the edit distance between the character string expressing the pronunciation of the candidate character string in the second language and the input character string is calculated. The similar word determination method according to claim 1, wherein: When the input character string consisting of the second language can be converted into the first language, the step of converting the input character string consisting of the second language into the first language;
Calculating an edit distance between the character string converted into the first language and the candidate character string;
Determining a candidate character string whose edit distance is not more than a reference value among the candidate character strings as a similar word of the input character string;
The similar word determination method according to claim 1, further comprising: In the step of converting to the first language,
When the first language is a foreign language, the input character string composed of the second language is converted into the character string composed of the corresponding foreign language by using a sound difference restorer for the corresponding foreign language. The foreign language similar word determination method according to claim 5.   The method of claim 1, wherein the first language is any one of foreign languages, and the second language is a native language.   The similar word determination method according to claim 1, wherein the input character string and the candidate character string are search queries. Prior to the determining step, further comprising receiving the input string from a user terminal;
The method of claim 1, further comprising the step of providing the determined similar word as a recommended query to the user terminal after the similar word determining step. The candidate string is
A candidate character string whose edit distance to the input character string is a reference value or less among candidate character strings stored in advance, or the input character string among candidate character strings including characters common to the input character string, The similar word determination method according to claim 1, wherein the character similarity score is selected from at least one of candidate character strings that are within the top N ranks.   The candidate character string whose edit distance is not more than a reference value among the candidate character strings is selected using a wild card character search for each calculation for calculating the edit distance. Similar word determination method of description. The candidate character string including characters common to the input character string is a candidate character string including N-gram common to the input character string,
The character similarity score includes the N-gram size common to the input character string, the number of the common N-grams, the similarity of the position where the common N-gram is found, and the input character string. The method according to claim 10, wherein the determination is performed using a difference in length between the candidate character strings and the candidate character strings.   A computer-readable recording medium having recorded thereon a program for executing the method according to any one of claims 1 to 12. A user interface unit that receives an input character string from a user terminal and provides the user terminal with a similar word for the input character string as a recommended query;
When the input character string is the second language, among the candidate character strings written in the first language, the character string expressing the pronunciation of the candidate character string in the second language, the input character string, A candidate character string written in the first language corresponding to a character string written in the second language whose edit distance is equal to or less than a reference value is determined as a similar word of the input character string and provided to the user interface unit A similar word determination unit;
Similar word determination system characterized by including. The similar word determination unit
An input character string determination unit for determining whether the input character string is in a first language or a second language;
When the input character string is in the second language, a first edit distance calculating unit that calculates an edit distance between the character string in which the pronunciation of the candidate character string is expressed in the second language and the input character string;
A first determination unit that determines a candidate character string expressed in the second language, the edit distance of which is less than or equal to a reference value, among the candidate character strings as a similar word of the input character string;
The similar word determination system according to claim 14, comprising: The similar word determination unit
A character string conversion necessity determination unit that determines whether or not the input character string written in the second language can be converted into a character string written in the first language;
When the first edit distance calculation unit cannot convert the character string composed of the second language into the first language, the character string expressing the pronunciation of the candidate character string in the second language and the input character string The similar word determination system according to claim 15, wherein the edit distance is calculated. When the second language is a native language,
15. The similar word determination system according to claim 14, wherein a character string in which the pronunciation of the candidate character string is expressed in the second language is converted by a sound difference converter for the native language. . The similar word determination unit
A second editing distance calculating unit that calculates an editing distance between the candidate character string and the input character string when the input character string is in the first language;
A second determination unit that determines a candidate character string whose edit distance is equal to or less than a reference value among the candidate character strings as a similar word of the input character string;
The similar word determination system according to claim 14, further comprising: The similar word determination unit
When the input character string consisting of the second language can be converted into the first language, the image processing apparatus further includes a character string conversion unit that converts the input character string consisting of the second language into the first language,
The similar word determination system according to claim 18, wherein the second editing distance calculation unit calculates an editing distance between the character string converted into the first language and the candidate character string. The character string converter
When the first language is a foreign language, it is a sound difference restorer for the corresponding foreign language that converts an input character string composed of the second language into a character string composed of the corresponding foreign language. Item 20. The foreign language similar word determination system according to Item 19.   The similar word determination system according to claim 14, wherein the first language is any one of foreign languages, and the second language is a native language.   The similar word determination system according to claim 14, wherein the input character string and the candidate character string are search queries.

Images