JP2007323250A - System, processing method and program for registering in dictionary for kana-kanji conversion - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a system, a processing method and a program for registering in a dictionary for KANA-KANJI conversion, capable of appropriately registering technical terms in the dictionary for KANA-KANJI conversion. <P>SOLUTION: The system 100 for registering the dictionary for KANA-KANJI conversion resisters technical terms in a technical term dictionary that is the dictionary for the KANA-KANJI conversion, and includes a technical term dictionary storage unit 10 that registers the technical terms to the technical terms dictionary which is a KANA-KANJI conversion dictionary and stores the technical term dictionary; a term origin generation unit 30 generating first and second origin expressing features of technical terms and non-technical terms and generating third and fourth origin expressing features of aspects of appearances of the technical terms and non-technical terms in a document; and a technical term determining unit 40 and a technical term registering unit 60 determining whether or not character strings in a new document are the technical terms on the basis of the first to the fourth origin to register the character strings determined as the technical terms in the technical term dictionary. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a kana-kanji conversion dictionary registration system for registering a term in a kana-kanji conversion dictionary used in creation of an electronic document, a processing method in the kana-kanji conversion dictionary registration system, and the kana-kanji conversion dictionary registration The present invention relates to a program executed in the system.

  When an electronic document with a high level of expertise is created, such as an electronic medical record created in the medical field, technical terms are often not registered in the dictionary for kana-kanji conversion. There is a problem that the time cost required increases. In order to solve this problem, it is necessary for a user, that is, a document writer, to register technical terms in the kana-kanji conversion dictionary in advance. If a necessary and sufficient amount of technical terms (and their readings) are registered in the Kana-Kanji conversion dictionary, for example, the corresponding conversion candidates can be displayed simply by entering the first three characters of the reading. Creation efficiency can be improved. However, especially when there are a large number of technical terms, such as in the medical field, it is difficult to comprehensively register technical terms in the kana-kanji conversion dictionary manually.

Regarding the registration of terms in the kana-kanji conversion dictionary, the technique described in Patent Document 1 is a system that reduces the burden of registration for each user by sharing a common kana-kanji conversion dictionary with a plurality of users. Has been proposed. Moreover, in the technique described in Patent Document 2, a system has been proposed in which the example of kana-kanji conversion performed by the user is automatically registered in the dictionary for kana-kanji conversion, thereby reducing the registration work by the user. .
Japanese Patent Laid-Open No. 11-66059 JP-A-9-134354

  However, even if the system in Patent Document 1 is adopted, it is still difficult to say that comprehensive registration of technical terms is possible, and furthermore, there is no situation in which a plurality of users jointly create documents in the same field. In some cases, the effect cannot be obtained. Further, in the system described in Patent Document 2, even terms that do not necessarily need to be registered may be registered in the kana-kanji conversion dictionary. As a result, a large number of unnecessary conversion candidates are displayed, and instead the document There arises a problem that the production efficiency is lowered.

  SUMMARY OF THE INVENTION An object of the present invention is to solve the above-mentioned problems, and a kana-kanji conversion dictionary registration system and a kana-kanji conversion dictionary registration process capable of appropriately registering technical terms in a kana-kanji conversion dictionary. A method and a dictionary registration program for kana-kanji conversion are provided.

  A kana-kanji conversion dictionary registration system for registering a term in a dictionary for kana-kanji conversion according to the present invention includes a storage means for storing a dictionary for kana-kanji conversion of technical terms, and a first feature representing characteristics of the technical terms First generation means for generating a second feature for generating a second feature representing a feature of an appearance of the technical term in the first document, and the first and second features And determining means for determining whether a character string of any one of the first document and the input second document different from the first document is a technical term, Registration means for registering a character string determined to be a term in the dictionary for kana-kanji conversion.

  According to this configuration, the feature of the technical term and the feature of the appearance of the technical term in the first document are analyzed, and based on these, whether the target character string is the technical term or not is determined. If it is determined that it is a technical term, it is registered in the dictionary for kana-kanji conversion so that it can be determined with high accuracy whether it is a technical term based on a previously created document. Thus, it becomes possible to appropriately register the technical terms in the kana-kanji conversion dictionary.

  In the dictionary registration system for kana-kanji conversion according to the present invention, the determination unit specifies a candidate for a technical term out of character strings other than the technical term in the first document, and stores the candidate in the second document. If the character string matches the candidate for the technical term, it may be determined that the character string is a technical term.

  According to this configuration, it is possible to register the target character string in the kana-kanji conversion dictionary more appropriately by registering the character string that matches the candidate technical term as the technical term in the kana-kanji conversion dictionary. Become.

  In the kana-kanji conversion dictionary registration system of the present invention, the first feature represents a part-of-speech sequence obtained by morphological analysis of the technical term, the number of characters of the technical term, and the content of katakana of the technical term. , The second feature is the frequency at which the terminology appears in the first document, the frequency at which the terminology modifies the verb by a predetermined case, and other sentences included in the sentence containing the terminology. You may make it represent the average value of the number of technical terms.

  In addition, the kana-kanji conversion dictionary registration system of the present invention includes a third generation unit that generates a third feature representing a character string characteristic other than the technical term in the first document, and the first generation unit. And a fourth generation unit that generates a fourth feature representing a feature of appearance of the character string other than the technical term in the document in the first document, and the determination unit includes the first feature Based on the third and fourth features, it may be determined whether the character string in the second document is a technical term.

  According to this configuration, it is possible to more appropriately determine whether the target character string is a technical term based on not only the positive example feature of the technical term but also the negative example feature.

  In the kana-kanji conversion dictionary registration system according to the present invention, the third feature is a part-of-speech sequence obtained by morphological analysis of a character string other than the technical term, the number of characters in the character string other than the technical term, It represents the content of katakana in a character string other than the term, the fourth feature is the frequency at which the character string other than the technical term appears in the first document, and the character string other than the technical term is a predetermined case. The frequency of modifying the verb may be represented by an average value of the number of character strings other than the technical terms included in the sentence including the character strings other than the technical terms.

  In the kana-kanji conversion dictionary registration system of the present invention, the kana-kanji conversion dictionary includes a character string selected by a predetermined operation instruction among the character strings determined by the registration means as the technical term. You may make it register to.

  According to this configuration, by making a final determination as to whether or not the user should register the target character string in the kana-kanji conversion dictionary, the target character string is more appropriately converted to kana-kanji conversion. It becomes possible to register in the dictionary for use.

  In the kana-kanji conversion dictionary registration system according to the present invention, a character string that is not registered in the kana-kanji conversion dictionary among the character strings determined by the third feature generation means as the technical term. For the character string that is not registered in the dictionary for kana-kanji conversion among the character strings that are determined to be the technical term, the fourth feature generation means generates the third feature, A fourth feature may be generated.

  According to this configuration, although it is determined to be a technical term, for example, a character string that is not instructed to be registered in the kana-kanji conversion dictionary by the user is treated as a negative term, that is, as a negative example, By causing the third and fourth features to be generated, in the subsequent determination of the technical term, the target character string is more appropriately converted into the technical term based on the characteristics of the positive and negative examples of the technical term. It can be determined whether or not there is.

  According to the present invention, a processing method in a kana-kanji conversion dictionary registration system for registering a term in a kana-kanji conversion dictionary generates a first feature representing a feature of a technical term included in the kana-kanji conversion dictionary. Based on the first and second features, a second generating step for generating a second feature that represents a feature of the appearance of the terminology in the first document, and A determination step of determining whether any one of the character strings in the first document and the second document is a technical term, and the character string determined to be the technical term as the dictionary for kana-kanji conversion And a registration step of registering to the system.

  Further, in the processing method of the present invention, the determination step specifies technical term candidates among character strings other than the technical terms in the first document, and the character strings in the second document are When matching with a candidate for a technical term, it may be determined that the character string is a technical term.

  The processing method of the present invention includes a third generation step of generating a third feature representing a character string characteristic other than the technical terms in the first document, and the special processing in the first document. A fourth generation step of generating a fourth feature representing a feature of the appearance of the character string other than the term in the first document, and the determination step includes the third and fourth steps. Based on the feature, it may be determined whether the character string in the second document is a technical term.

  The program executed in the kana-kanji conversion dictionary registration system for registering a term in the kana-kanji conversion dictionary according to the present invention generates a first feature representing the characteristics of the technical terms included in the kana-kanji conversion dictionary. A first generation step, a second generation step for generating a second feature representing a feature of the appearance of the technical term in the first document, and the first and second features. A determination step of determining whether any one of the character strings in the first document and the second document is a technical term, and converting the character string determined to be the technical term into the kana-kanji conversion And a registration step of registering in the dictionary for use.

  Further, in the program according to the present invention, the determining step specifies a candidate for a technical term among character strings other than the technical term in the first document, and the character string in the second document is the specialized text. When matching with a term candidate, the character string may be determined to be a technical term.

  The program of the present invention includes a third generation step of generating a third feature representing a character string characteristic other than the technical term in the first document, and the technical term in the first document. And a fourth generation step for generating a fourth feature representing the feature of the appearance mode of the character string other than the first character in the first document, and the determination step includes the third and fourth features. Based on the above, it may be determined whether the character string in the second document is a technical term.

  According to the present invention, by determining whether a target character string is a technical term based on characteristics of the technical term and characteristics of an appearance mode of the technical term in the document, It becomes possible to appropriately register in the kana-kanji conversion dictionary.

  Embodiments of the present invention will be specifically described with reference to the drawings. FIG. 1 is a diagram showing a configuration example of a dictionary registration system for kana-kanji conversion. The kana-kanji conversion dictionary registration system 100 shown in the figure is configured by a single information processing apparatus such as a personal computer, and identifies technical terms among terms included in an electronic document, for kana-kanji conversion. Register in the dictionary.

  This kana-kanji conversion dictionary registration system 100 includes a technical term dictionary storage unit 10 corresponding to a storage unit, a document storage unit 20, a term feature generation unit 30 corresponding to first to fourth generation units, and a determination unit. The technical term determination unit 40, the technical term candidate storage unit 50, the technical term registration unit 60 corresponding to the determination unit and the registration unit, the input unit 70, and the monitor 80. Among these, the technical term dictionary storage unit 10, the document storage unit 20, the term feature generation unit 30, the technical term determination unit 40, the technical term candidate storage unit 50, and the technical term registration unit 60 are configured by a CPU and a memory. Is realized by executing a predetermined program.

  The technical term dictionary storage unit 10 stores a kana-kanji conversion dictionary (technical term dictionary) in which technical terms in the medical field are registered. In the technical term dictionary, a character string of a technical term and a reading of the technical term are registered in association with each other. The document storage unit 20 is created using a technical term dictionary, and stores a set of electronic documents including the technical term.

  The term feature generation unit 30 includes a first feature that represents a feature of the technical term for the technical term registered in the technical term dictionary stored in the technical term dictionary registration unit 10, and the technical term is a document storage unit. A second feature is generated that indicates how the document set is appearing in the document set that is the set of first documents stored in 20.

  The technical term determination unit 40 is a technical term registered in the technical term dictionary in the document set stored in the document storage unit 20 based on the first and second features generated by the term feature generation unit 30. For any character string other than, it is determined whether there is a possibility of a technical term. The technical term candidate storage unit 50 stores a character string determined to be a technical term as a technical term candidate.

  The technical term registration unit 60 determines whether an arbitrary character string in the document created by the user received by the input unit 70 is a technical term based on the technical term candidate stored in the technical term candidate storage unit 50. Determine. When the character string is a technical term, the technical term registration unit 60 adds the technical term and the reading of the technical term in association with each other to the technical term dictionary stored in the technical term dictionary storage unit 10.

  The detailed operation of the kana-kanji conversion dictionary registration system 100 will be described below with reference to the flowchart. FIG. 2 is a flowchart showing the extraction operation of technical term candidates of the dictionary registration system 100 for kana-kanji conversion.

  The term feature generation unit 30, for the technical terms registered in the technical term dictionary stored in the technical term dictionary registration unit 10, the part-of-speech string obtained by morphological analysis of the technical terms, the number of characters of the technical terms, the katakana of the technical terms The first feature consisting of the percentage of content, the frequency with which technical terms appear in the document set, and the special terms in the document set qualify specific verbs as wo case, second case, de case, ga case, etc. A second feature comprising an average value of the frequency and the number of other technical terms existing in the sentence including technical terms in the document set is generated (S101).

  For example, when the technical term is “α lipoic acid”, the part-of-speech sequence obtained by morphological analysis of the technical term in the first feature is “Greek letter-Katakana-suffix”, and the technical term has 4 characters. Yes, the katakana content of the technical term is 2/4 = 50%. If the technical term is “α lipoic acid” and there are 47 expressions “prescribe α lipoic acid” in the document set, the technical term in the second feature is a verb in the document set. The frequency of modifying “Prescribe” with wo is 47. These first and second features are used as positive examples in specifying technical term candidates in the technical term determination unit 40 described later.

  Next, the term feature generation unit 30 specifies a character string that is not registered in the technical term dictionary from the document set stored in the document storage unit 20 based on the first and second features, The compound noun A is extracted from the character string (S102). Further, the term feature generation unit 30 rearranges the extracted compound nouns A in the order of appearance frequency in the document set stored in the document storage unit 20, and extracts the top N items having the lower frequency (S103). ).

  Next, the term feature generation unit 30 extracts, for the extracted N compound nouns A, the third feature representing the characteristics of the compound noun A and the document in which the compound noun A is stored in the document storage unit 20. A fourth feature is generated that indicates what kind of aspect appears in the set (S104). Specifically, the third feature includes a part-of-speech sequence obtained by morphological analysis of the compound noun A, the number of characters of the compound noun A, and the content of katakana of the compound noun A, as in the first feature. The feature of, like the second feature, is the frequency with which the compound noun A appears in the document set, and the compound noun A modifies the specific verb in the document set with wo case, second case, de case, ga case, etc. Frequency, and an average value of the number of other compound nouns existing in a sentence including compound nouns in a document set. These third and fourth features are used as negative examples in specifying technical term candidates in the technical term determination unit 40 described later.

  The first to fourth features generated by the term feature generation unit 30 are sent to the technical term determination unit 40 together with the extracted N compound nouns A.

  For each of the N compound nouns A, the technical term determination unit 40 creates a technical term dictionary based on the first and second features that are positive examples and the third and fourth features that are negative examples. It is determined whether there is a possibility of a technical term to be registered (S105).

  Here, a technique based on SVM (Support Vector Machine) learning (Vapnik, V. N., “The Nature of Statistical Learning Theory, Springer”, 1995) can be used as a representative determination technique. In the determination method based on SVM learning, the first and second features of the positive example and the third and fourth features of the negative example are used as learning data, and learning for discriminating between the positive example and the negative example can be performed. Become. Therefore, it is possible to determine whether an arbitrary character string is a technical term or a non-technical term based on the first to fourth features, and the tendency of registration of the technical term in the user's technical term dictionary is increased. Along with this, it is possible to distinguish between technical terms and non-technical terms.

  Further, the technical term determination unit 40 stores the compound noun A determined as having a high possibility of the technical term in the technical term candidate storage unit 50 as the technical term candidate B (S106).

  FIG. 3 is a flowchart showing the registration operation of technical terms in the dictionary registration system 100 for kana-kanji conversion.

  In order to create a document, the user operates an operation unit (not shown) such as a keyboard to input Japanese text. The input unit 70 has a user interface and creates a (second document) document according to the operation of the operation unit by the user. A technical term dictionary stored in the technical term dictionary storage unit 10 is used for kana-kanji conversion at the time of document creation. The created document is sent to the technical term registration unit 60.

  The technical term registration unit 60 inputs the document from the input unit 70 (S201). Next, the technical term registration unit 60 specifies a compound noun C that is a character string included in the input document (S202), and the compound noun C is stored in the technical term candidate storage unit 50. It is determined whether or not it matches any one (S203).

  If the compound noun C does not match any of the technical term candidates B, the series of operations ends. On the other hand, when the compound noun C matches any one of the technical term candidates B, the technical term registration unit 60 determines that the compound noun C is a technical term (S204), and the technical term is referred to the technical term dictionary for the user. A screen (registration instruction screen) for prompting registration is displayed on the monitor 80 (S205).

  FIG. 4 is a diagram illustrating an example of a registration instruction screen. The registration instruction screen shown in FIG. 4 includes a document 302 from the input unit 70 and a display 302 for selecting technical term registration. Further, the display 302 for selecting the registration includes a button 303 displayed as “Yes” selected when registering the technical term, and “No” selected when not registering the technical term. ”Is displayed.

  Next, the technical term registration unit 60 determines whether or not registration of technical terms is instructed by the user (S206). For example, when the user operates an operation unit (not shown) such as a mouse and selects the button 303 displayed as “Yes” on the registration instruction screen shown in FIG. 4, the technical term registration unit 60 displays the technical term If it is determined that registration has been instructed and the button 304 displayed as “No” is selected, the technical term registration unit 60 determines that it has been instructed not to register the technical term.

  When it is instructed not to register a technical term, a series of operations is completed. On the other hand, when the registration of the technical term is instructed, the technical term registration unit 60 reads the compound noun C that is the technical term and the reading of the compound noun C that is input according to the operation of the operation unit of the user. Correspondingly, it is registered in the technical term dictionary stored in the technical term dictionary storage unit 10 (S207).

  Thereafter, the created document is stored in the document storage unit 20, and the operation shown in FIG. 2 is further performed to update the technical term candidates stored in the technical term candidate storage unit 50. Further, every time a document created by the user is input from the input unit 70, the operation shown in FIG. 3 is performed, and the technical term dictionary stored in the technical term dictionary storage unit 10 is updated.

  FIG. 5 is a flowchart showing another operation for registering technical terms in the dictionary registration system 100 for kana-kanji conversion. The operations in S301 through S307 in FIG. 5 are the same as the operations in S301 through S307 in FIG.

  In S306, when it is determined that it is instructed not to register the compound noun C determined to be a technical term, the technical term determination unit 40 recognizes that the compound noun C is not a technical term, and N N It adds to compound noun A (S308).

  Here, the technical term determination unit 40 has a boosting function. Boosting is an error-driven machine learning method. When a case where the learning result is determined to be a positive example is actually a negative example, in order to correct the determination result of that case correctly, the learning data of that case (See Hideki Aso, Koji Tsuda, Noboru Murata, “Statistics of Pattern Recognition and Learning”, Iwanami Shoten, 2003). The technical term determining unit 40 regards that the compound noun C is not a technical term and adds it to the N compound nouns A as an instruction not to register the compound noun C as a technical term. Thereafter, the term feature generation unit 30 generates the third and fourth features that are negative examples for the compound noun C. Thereby, in determination of a technical term after that, it can be determined more appropriately whether a target character string is a technical term in consideration of a user's judgment.

  As described above, the kana-kanji conversion dictionary registration system 100 has the first feature that is the feature of the technical term in the document created in the past, and the second feature that represents the feature of the appearance of the technical term in the document. Newly created based on the third feature that is the feature of the non-technical term in the document created in the past, and the fourth feature that represents the feature of the appearance of the non-technical term in the document. It is based on documents created in the past by determining whether or not the character string in the document is a technical term and registering it in the technical term dictionary if it is determined to be a technical term. Thus, it is possible to determine whether the term is a technical term with high accuracy, and it is possible to appropriately register the technical term in the technical term dictionary.

  In the above-described embodiment, the kana-kanji conversion dictionary registration system 100 is configured in a single information processing apparatus such as a personal computer, but is configured by a plurality of information processing apparatuses connected via a network. May be.

  FIG. 5 is a diagram showing another configuration of the kana-kanji conversion dictionary registration system. A kana-kanji conversion dictionary registration system 500 shown in FIG. 5 includes a plurality of personal computers (PCs) 200 and a server 250 connected to a network 400, and a technical term dictionary database (DB) 260 managed by the server 250. Composed.

  The PC 100 has the same configuration as that obtained by removing the technical term dictionary storage unit 10 from the kana-kanji conversion dictionary registration system 100 described above. When the PC 100 generates the first and second features, the PC 100 requests the technical term dictionary stored in the technical term dictionary DB 260 from the server 250. In response to this request, the server 250 reads the technical term dictionary stored in the technical term dictionary DB 260 and sends it to the PC 100.

  In addition, when the technical term is specified, the PC 100 associates the technical term and the reading of the technical term and transmits the technical term to the server 250 via the network 400. The server 250 registers the received technical term and the reading of the technical term in the technical term dictionary stored in the technical term dictionary DB 260.

  As a result, the technical term dictionary can be shared by a plurality of users, and it can be determined with higher accuracy whether or not it is a technical term.

  Further, in the embodiment described above, a candidate for a technical term is specified, a character string in a new document that matches the technical term is determined to be a technical term, and only when registration is instructed by the user Although the character string is registered in the technical term dictionary, the technical term candidates may be registered in the technical term dictionary as they are. In this case, the burden on the user can be reduced and new technical terms can be automatically registered in the technical term dictionary.

  As described above, the kana-kanji conversion dictionary registration system, the kana-kanji conversion dictionary registration processing method, and the kana-kanji conversion dictionary registration program according to the present invention appropriately register technical terms in the kana-kanji conversion dictionary. This is useful as a dictionary registration system for kana-kanji conversion.

It is a figure which shows the structure of the dictionary registration system for kana-kanji conversion. It is a flowchart which shows operation | movement of the technical term candidate specification of the dictionary registration system for kana-kanji conversion. It is a flowchart which shows operation | movement of the technical term registration of the dictionary registration system for kana-kanji conversion. It is a figure which shows an example of a registration instruction | indication screen. It is a flowchart which shows other operation | movement of the technical term registration of the dictionary registration system for kana-kanji conversion. It is a figure which shows the other structure of the dictionary registration system for kana-kanji conversion.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 Technical term dictionary memory | storage part 20 Document memory | storage part 30 Term feature generation part 40 Technical term judgment part 50 Technical term candidate memory | storage part 60 Technical term registration part 70 Input part 80 Monitor 100 Kana / Kanji conversion dictionary registration system

Claims (13)

A kana-kanji conversion dictionary registration system for registering terms in a kana-kanji conversion dictionary,
Storage means for storing a dictionary for conversion of technical terms to kana-kanji;
First generating means for generating a first feature representing the feature of the technical term;
Second generating means for generating a second feature representing a feature of an appearance mode of the technical term in the first document;
Based on the first and second features, whether or not any character string of the first document and the input second document different from the first document is a technical term Determination means for determining;
A kana-kanji conversion dictionary registration system comprising: a registering unit that registers a character string determined to be the technical term in the kana-kanji conversion dictionary.   The determination unit identifies a technical term candidate among character strings other than the technical term in the first document, and the character string in the second document matches the technical term candidate. The kana-kanji conversion dictionary registration system according to claim 1, wherein the character string is determined to be a technical term. The first feature represents a part-of-speech string obtained by morphological analysis of the technical term, the number of characters of the technical term, and the content of katakana of the technical term,
The second feature includes the frequency with which the technical term appears in the first document, the frequency with which the technical term modifies the verb according to a predetermined case, and other specialties included in the sentence containing the technical term. The kana-kanji conversion dictionary registration system according to claim 1 or 2, wherein an average value of the number of terms is expressed. Third generation means for generating a third feature representing a character string characteristic other than the technical term in the first document;
And fourth generation means for generating a fourth feature representing a feature of an appearance mode of the character string other than the technical term in the first document in the first document,
4. The determination unit according to claim 1, wherein the determination unit determines whether the character string in the second document is a technical term based on the third and fourth features. The dictionary registration system for kana-kanji conversion described in 1. The third feature represents a part-of-speech sequence obtained by morphological analysis of a character string other than the technical term, the number of characters of the character string other than the technical term, and the content of katakana in the character string other than the technical term,
The fourth feature is a frequency at which a character string other than the technical term appears in the first document, a frequency at which a character string other than the technical term modifies a verb according to a predetermined case, and other than the technical term 5. The dictionary registration system for kana-kanji conversion according to claim 4, representing an average value of the number of character strings other than other technical terms included in a sentence including the character string.   6. The registration means for registering a character string selected by a predetermined operation instruction among character strings determined to be the technical term in the kana-kanji conversion dictionary. A dictionary registration system for kana-kanji conversion according to any one of the above. The third feature generation means generates the third feature for a character string that is not registered in the kana-kanji conversion dictionary among character strings determined to be the technical term,
The fourth feature generation means generates the fourth feature for a character string that is not registered in the dictionary for kana-kanji conversion among character strings determined to be the technical term. The kana-kanji conversion dictionary registration system according to claim 4 or 5. A processing method in a dictionary registration system for kana-kanji conversion that registers terms in a dictionary for kana-kanji conversion,
A first generation step of generating a first feature representing characteristics of technical terms included in the dictionary for kana-kanji conversion;
A second generating step for generating a second feature representing a feature of the appearance of the terminology in the first document;
A determination step of determining whether any one of the character strings in the first document and the second document is a technical term based on the first and second features;
And a registration step of registering a character string determined to be the technical term in the kana-kanji conversion dictionary.   The determination step specifies a technical term candidate from character strings other than the technical term in the first document, and the character string in the second document matches the technical term candidate. The processing method according to claim 8, wherein the character string is determined to be a technical term. A third generation step of generating a third feature representing characteristics of a character string other than the technical terms in the first document;
A fourth generation step of generating a fourth feature representing a feature of an appearance mode of the character string other than the technical term in the first document in the first document;
9. The determination according to claim 7, wherein the determination step determines whether or not the character string in the second document is a technical term based on the third and fourth features. Processing method. A program executed in a kana-kanji conversion dictionary registration system for registering terms in a kana-kanji conversion dictionary,
A first generation step of generating a first feature representing characteristics of technical terms included in the dictionary for kana-kanji conversion;
A second generating step for generating a second feature representing a feature of the appearance of the terminology in the first document;
A determination step of determining whether any one of the character strings in the first document and the second document is a technical term based on the first and second features;
And a registration step of registering a character string determined to be the technical term in the kana-kanji conversion dictionary.   The determination step specifies a technical term candidate from character strings other than the technical term in the first document, and the character string in the second document matches the technical term candidate. The program according to claim 11, wherein the character string is determined to be a technical term. A third generation step of generating a third feature representing characteristics of a character string other than the technical terms in the first document;
A fourth generation step of generating a fourth feature representing a feature of an appearance mode of the character string other than the technical term in the first document in the first document;
The said determination step determines whether the character string in the said 2nd document is a technical term based on the said 3rd and 4th feature, It is characterized by the above-mentioned. program.

Images