JPH10134060A - Equivalent generating method, machine translation system for executing the method, and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain an equivalent out of one proper technical term dictionary by extracting a relative technical field having close relation to some technical field, adding and evaluating relative points by the technical field and relative technical fields, and selecting the equivalent out of the technical term dictionary having high evaluation. SOLUTION: When a language to be translated is inputted after environment data on a users name, a technical field name, etc., including technical terms are inputted, a translating process control part 1 stores the environment data in an environment data storage part 8 and then counts the frequency of combined use of a user dictionary and a technical term dictionary. According to the use frequency, relation points set in a user dictionary relation table 9 are updated by an update control means 21. On the basis of the relation points after the update by this update control means 21, an equivalent generation control means 22 generates an equivalent. When one equivalent is selected and generated, the selection result of the equivalent is learnt by a learning control means 23.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a novel translated word generation method, a machine translation system for executing the translated word generation method, and a recording medium.

[0002]

2. Description of the Related Art Conventional machine translation systems include standard dictionaries for defining translations of general languages, a plurality of user dictionaries for defining translations available for each user, and specialized fields such as electrical, mechanical, and chemical. There are provided a plurality of technical term dictionaries that define the translation of each term, and a learning dictionary that stores learning results. In each dictionary, a headword and a translation corresponding to the headword are specified, and at least one or more translations are prepared for each headword, and priority is set between the translations. ing.

[0003] In actual translation, after setting a translation field in a particular field by a certain user, when a language to be translated is input, a user dictionary and a technical term dictionary corresponding to the set contents are selected. It finds a translation from the dictionary entry, translates it into the target language, and translates it.

By the way, in the above-described machine translation system, when a translation word for a language to be translated is selected, learning and storage is performed in a learning dictionary. At this time, a combination of each user dictionary and each technical term dictionary is used. It is a method of learning and storing.

[0005]

Therefore, in the above-described machine translation system, learning and storing are performed for each combination of each user dictionary and each technical term dictionary. In addition, since the number of combined patterns becomes enormous and cannot be used for any other combinations, there is a disadvantage that the learning result cannot be used effectively.

[0006] In addition, there is a possibility that a plurality of translations exist in a plurality of dictionaries for a certain headword. In this case, there is a problem in which translation is output as the first translation. In many cases, a dictionary to be searched is determined based on the priority order among the dictionaries, and a high-priority translation within the corresponding dictionary is selected and output, but the search process takes a lot of time, and only the priority order is used. There is a problem that it is not possible to select an accurate translation from selecting a translation.

[0007] The invention described in claim 1 and claim 2 is a method for translating a language of a specialized field from an appropriate technical term dictionary while considering a technical term dictionary of a related specialized field related to the specialized field. An object of the present invention is to provide a translated word generation method for translating a translated word.

[0008] The invention described in claims 3 to 5 provides a machine translation system that specifies a highly-recognized technical term dictionary for translation of a language in a certain specialty field and translates an appropriate translation. It is in.

The invention described in claims 6 to 8 allows a computer to read a program recorded in a storage medium and execute the program, thereby considering a technical term dictionary of a related specialty field related to the specialty field. Another object of the present invention is to translate an appropriate translated word from one appropriate technical term dictionary.

[0010]

The invention corresponding to claims 1 and 2 is a method of converting a technical term dictionary into a score according to the degree of closeness between specialized fields and receiving a translation instruction of a language in a specific field. From the points in the relation table, extract the related specialty field that has a close relationship with the particular specialty field, and when there is a translation result in the technical term dictionary of the relevant specialty field, the relevant score for each of the certain specialty field and the related specialty field This is a translated word generation method that adds and evaluates, specifies a specialized term dictionary in a highly evaluated specialized field, and translates a translated word from the specialized term dictionary.

According to the inventions corresponding to the first and second aspects of the present invention, when the above-mentioned measures are taken and a translation instruction of a language in a certain specialty field is received, a related specialty field closely related to the certain specialty field is obtained. And for each of these specializations, the related score is added and evaluated based on the score indicating the closeness between the specializations. For example, one specialty having the highest evaluation score can be specified. Translated words can be translated from the technical term dictionary, and appropriate translated words can be translated without complicated search processing steps.

Next, a third aspect of the present invention relates to a technical term dictionary for defining a translation word of each technical field, and a technical term dictionary for storing a score in accordance with the closeness between the respective technical fields. Table, when receiving a translation instruction of a language of a certain specialty field, extract a related specialty field having a close relationship with the certain specialty field from the score of the technical term dictionary relation table, and extract a technical term dictionary of the relevant specialty field When there is a translation result, the related score is added and evaluated for each of the certain specialty field and the related specialty field, and the technical term dictionary of one specialty field is specified from the highest evaluation score, and from the technical term dictionary, This is a machine translation system provided with a translated word generation control means for translating translated words.

According to the third and fifth aspects of the present invention, when a translation instruction of a language of a certain specialty field is received by taking the above-described means, a related specialty field closely related to the certain specialty field is obtained. Judge whether or not there is a translation result of the same language in the technical term dictionary, if there is a translation result, for a certain specialty field and a related specialty field with the translation result, the related score from the score of the technical term dictionary relation table Extraction and addition evaluation are performed, and the terminology dictionary of a certain specialty field is specified from the evaluation score with the highest evaluation, so that an appropriate translated word can be translated.

According to a fourth aspect of the present invention, there is provided a technical term dictionary for defining a translation of each technical field, a plurality of auxiliary dictionaries for defining a translation of an auxiliary field necessary for translating the translation of the specialized field, A technical term dictionary relation table that stores the degree of closeness between each specialty field as a score, an auxiliary dictionary relation table that stores a score that changes according to the number of uses of the auxiliary field and the specialty field, An update control means for counting the number of times of use of the certain auxiliary field and the specialty field when receiving a language translation instruction in relation to the specialty field, and updating the score of the auxiliary dictionary relation table; Means for determining whether a translation instruction is for a language based on an auxiliary field and a specialty field or a language translation instruction based on another auxiliary field and a specialty field; To When a word translation instruction is received, a related specialty field having a close relationship with the certain specialty field is extracted from the scores of the technical term dictionary relation table, and the related score is added for each of the certain specialty field and the related specialty field. Evaluate and translate a translation from one jargon dictionary, and receive a language translation instruction under the relationship between the other linguistic fields and the linguistic field, enter the linguistic dictionary relation table and the linguistic dictionary relation table. This is a machine translation system provided with a related word score added from a memorized score, and a translated word generation control means for translating a translated word from one highly evaluated attached dictionary and one technical term dictionary.

According to the fourth and fifth aspects of the present invention, by taking the above-described measures, an auxiliary dictionary of an auxiliary field that spans another area (for example, an auxiliary field in which the user “Sato” is the user “Suzuki”) is created. When a language translation instruction is received under the relationship between a specialized field and a specialized field, the related points are added and evaluated from the points stored in the technical term dictionary relation table and the auxiliary dictionary relation table. Since the translation is translated from the attached dictionary and one technical term dictionary, a more appropriate translation can be translated.

[0016] The invention corresponding to claim 6 is a technical term dictionary for defining a translation word of each technical field, means for scoring according to the closeness between the respective technical fields, and storing the points in the first storage means, When receiving a translation instruction of a language of a certain specialty field, a related specialty field having a close relationship with a certain specialty field is extracted from the score of the first storage means, and a relevant specialty field and a related specialty field are associated with each other. This is a computer-readable recording medium that stores a program that implements a means for adding and evaluating points and translating a translated word from a highly evaluated technical term dictionary.

By taking such means, when a translation instruction of a language of a certain specialty field is received, a related specialty field closely related to the certain specialty field is extracted from the score of the first storage means, Since the related score is added and evaluated for each of the certain specialty field and the related specialty field, and the translated word is translated from the highly evaluated technical term dictionary, an appropriate translated word can be translated.

Next, an invention corresponding to claim 7 is a technical term dictionary for defining translations for each specialty field, and a plurality of auxiliary dictionaries for defining translations for auxiliary fields necessary for translating the translations for the specialty fields. Means for storing scores in the first storage means in accordance with the closeness between the respective specialty fields, and second storage means for storing scores which change in accordance with the number of uses of the auxiliary fields and the specialty fields Receiving a language translation instruction in relation to a certain auxiliary field and specialty field, counts the number of times the certain auxiliary field and specialty field are used, and updates the score of the second storage means. Control means; means for determining whether a translation instruction for a language is based on a certain auxiliary field and specialty field or a translation instruction for a language based on another auxiliary field and specialty field; Language translation in relation to specialty When the instruction is received, a related specialty field having a close relationship with the certain specialty field is extracted from the score of the first storage unit, and the related score is added and evaluated for each of the certain specialty field and the related specialty field, When a translated word is translated from one technical term dictionary and a language translation instruction is received under the relation between the other specialty field and the specialty field, it is stored in the first storage means and the second storage means. A computer-readable recording medium that records a program for realizing a translated word generation control unit that translates a translated word from a highly evaluated one attached dictionary and one technical term dictionary by adding and evaluating a related score from the score.

By taking such a measure, when a language translation instruction is received in relation to a certain auxiliary field and a specialty field, the first storage means makes a close contact with the certain specialty field. Extract related specialty fields with relevant relationships, add and evaluate the relevant score for each of the certain specialty fields and the related specialty fields, translate words from one technical term dictionary, and When a translation instruction of a language is received under the relation of (1), the related points are added and evaluated from the points stored in the first storage means and the second storage means. Since the translation is translated from the technical term dictionary, the translation can be translated appropriately.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A translated word generation method according to the present invention will be described. First, when a translated language including a technical term is translated, a score is previously scored in a technical term dictionary relation table in accordance with the closeness between specialized fields such as electricity, information, electronics, machinery, architecture, and chemistry. For example, regarding electricity, electric-electron = 10, electric-information = 10, electric-machine = 5,
It is defined as follows: electricity-building = 1, electricity-chemistry = 5. Similarly, for information, electronics, machinery, architecture, chemistry, etc., the closeness between them is scored. Then, when receiving a translation instruction of the language to be translated, which is a technical term related to a certain technical field, for example, `` electricity, '' for example, `` information, '' which has a close relationship with the specific technical field from the score of the technical term dictionary relation table,
Extract the related specialty field of "electronic" and, when there is a translation result of the translated language that is the technical term in the technical term dictionary of the related specialty field, sequentially calculate the relevant score for each of the certain specialty field and the related specialty field The addition is performed, for example, to specify a technical term dictionary of a technical field having the largest value of the added points, and to translate a translation of the translated language, which is the technical term, from the technical term dictionary.

Next, FIG. 1 is a block diagram showing an embodiment of a machine translation system for carrying out the above-described method. In FIG. 1, reference numeral 1 denotes a translation processing control unit including a CPU for executing a predetermined process in accordance with a translation program, and includes a headword corresponding to a general language and at least one headword corresponding to the headword. One or more standard dictionaries 2 that define one or more translations, one or more user dictionaries that define headwords for each user such as an individual or a group, and at least one or more translations corresponding to the headwords ( Attached dictionary) 3 and one or more technical term dictionaries 4 that define headwords and at least one or more translations corresponding to the headwords for each specialized field such as electricity, information, machinery, architecture, and chemistry. A learning dictionary 5 for learning and storing translated word selection results for the translated language to be translated and a translation rule storage unit 6 for defining the priority of documents and each dictionary are provided. For the dictionaries 2 to 5 and the storage unit 6, an external storage device such as a hard disk or a floppy disk or a main storage device is used.

In this machine translation system, input devices such as a keyboard, a mouse, and the like for inputting environment data such as a user name and a specialty field name as a pre-translation operation, and for inputting a translated language and other necessary instruction data. 7, an environment data storage unit 8 for storing environment data, a user dictionary relation table 9, a technical term dictionary relation table 10, a display device 11 for displaying translation results and the like, and a printing device 12 for printing out translation results and the like are provided. Have been.

The technical term dictionary relation table 10 is a table expressing the closeness between a plurality of technical term dictionaries 3,... By scores. For example, the scores are tabulated as follows. However, for convenience, the score is described in a range not exceeding a maximum of 10 points.

Electric-Electronic 10 Electric-Information 10 Electric-Machine 5 Electric-Architecture 1 Electric-Chemical 5 Electron-Information 10 Electro-Machine 5 Electron-Architecture 1 Electronic-Chemical 5 Information-Machine 5 Information-Architecture 1 Information-Chemical 1 Machine-Architecture 5 Machine-Chemistry 1 Architecture-Chemistry 1 etc. In addition, there are many specialized fields in addition to the above, and for example, when the number of fields increases, a maximum of 1
The value may be set to exceed zero or more points.

The user dictionary relation table 9 is a table for defining the points while increasing and varying the points according to the number of times the combination of the user dictionary 3 and the technical term dictionary 4 is used, and decreasing and varying the points as necessary. For example, a user “Sato” has his own “Sato” user dictionary and “Suzuki”
When the user dictionary is used and used in combination with the “electric” and “electronic” terminology dictionaries, the user dictionary relation table 9 describes the following scores.

Sato-Suzuki 1 Sato-Electric 1 Sato-Electronic 1 This combination is also set, for example, within a range not exceeding 10 points for convenience. In this system, the more the combination of the same user dictionary and the terminology dictionary is used, the more points are used, for example, up to 10 points, and the number of times of use is reduced by decreasing the points for unused combinations. Set to appear in the score appropriately.

Next, the translation processing control unit 1 functionally
After inputting environmental data such as a user name and a technical field name including a technical term, and inputting a language to be translated, the environmental data is stored in the environmental data storage unit 8, and a combination of a user dictionary and a technical term dictionary is stored. Count the number of uses,
Update control means 21 for updating the number of relation points set in the user dictionary relation table 9 according to the number of times of use, translation word generation control means 22 for generating a translation based on the number of relation points updated by the update control means 21; The learning control means 23 learns the result of selecting a translation when a translation is selected and generated.

Next, the operation of the system configured as described above will be described. In this machine translation system, an environment such as a user name (user dictionary) “A” and a specialized field (technical term dictionary) “information” input as a pre-translation operation in the update processing unit 21 in the translation processing control unit 1. After the data is stored in the environmental data storage unit 8 (ST1), the number of combinations of the user name (user dictionary) and the technical field (technical term dictionary), which are the contents of the environmental data, is counted, and based on this count number. The number of relation points in the user dictionary relation table 9 is updated (ST2).

As the score updating process, for example, the following process is executed. As a pre-translation operation, environmental data of the user name and the specialty field name are stored in the environmental data storage unit 8, and the number of combinations of the user dictionary corresponding to the user name and the technical term dictionary corresponding to the specialty field name is counted. Each time the count reaches a predetermined count, the score is incremented and updated by one point. Also, a predetermined period is determined in advance, and a pre-translation operation for a combination of the same user dictionary and technical term dictionary is performed within this period. When there is no point, the points are updated by decreasing and updating the points one by one.

As another example, a combination of a user dictionary and a technical term dictionary and the combination date and time data are stored, a predetermined period is determined in advance, and the period between the current time and the combination date and time data is set to a predetermined period. When the count has elapsed, -1 is subtracted from the total number of counts. As a result, when the count has decreased to a predetermined count, an update process of decreasing the score by -1 point is performed, but various other update processes are conceivable.

After performing the updating process as described above, a translated word is generated by the translated word generation control means 22. The translated word generation control unit 22 learns the learning result of the input translated language by using the same combination of the user dictionary “A” and the technical term dictionary “information” in the learning dictionary 5 as before. It is checked whether or not it is stored (ST3). If there is the same learning result, the translated word is translated (ST4).

If there is no learning result, "electricity" and "electronic" which are related specialty fields closely related to "information" are extracted from the technical term dictionary relation table 10, and these "electricity" and "electronic" are extracted. It is determined whether there is a similar translation result in the technical term dictionary 4 (ST5). If there is no similar translation result in the technical term dictionary 4 of these related specialized fields “Electricity” and “Electronic”, a translation term is selected from the user dictionary 3 of “A” and the technical term dictionary 4 of “Information” (ST
6) The translated word translated by the learning control means 23 is stored in the learning dictionary 5
Is learned and stored (ST7).

On the other hand, if there is a similar translation result in the technical term dictionary 4 of “electric” and “electronic”, “information”, “electric”,
A score relating to “Electronic” is calculated (ST8), and the translated word result of the technical term dictionary 4 having the largest score as the calculation result is selected as the first translated word (ST9). Next, for example, when the user A inputs “information” with the users A and B as environmental data, the technical term dictionary 4 having a close relationship with “information” is obtained from the technical term dictionary relation table 10 in the same manner as described above.
Then, the related specialized fields of “Electricity” and “Electronic” are extracted, and it is determined whether or not there is a similar translation result in the technical term dictionary 4 of “Electricity” and “Electronic”. If there is a similar translation result in the technical term dictionary 4 for “electricity” and “electronic”, the user A
"Information", "Electricity", "Electronics"
Is calculated, and the translated word result of the technical term dictionary 4 having the largest calculated result is selected as the first translated word.
If there is no similar translation result in the technical term dictionary 4 of “Electricity” and “Electronic”, a translation is selected and translated from the user dictionary 3 of “A” and the technical term dictionary 4 of “Information”. Then, the translated word translated by the learning control means 23 is learned and stored in the learning dictionary 5.

Next, as a specific example, for example, when the user “Sato” uses the user dictionary 3 of the user “Suzuki” and the technical term dictionary of the technical terms “electronic” and “information”, “t”
The procedure for determining the translation of “ool” will be described.

At this stage, for example, "electricity"
Using the terminology dictionary 4 for “information” and “machine” and the user dictionary 3 for “Sato”, the “tool” is learned and stored as the first translation for the translation of “tool”.
Using the technical term dictionary 4 for “machine” and “architecture” and the user dictionary 3 for “Sato”, learn and memorize “tool” as the first translation for the translation of “tool”, and further, Using the dictionary 4 and the user dictionary 3 of "Suzuki",
It is assumed that “tool” is learning and storing the translated word “tool” as the first translated word.

As a result, the technical term dictionary relation table 10
Has a fixed number of relation points, but the attached dictionary relation table 9
It is assumed that the following points are described by the update control means 21 in the.

Sato-Electronics 10 Sato-Electric 10 Sato-Machinery 7 Sato-Suzuki 3 Sato-Takahashi 3 Suzuki-Information 10 Suzuki-Architecture 4 Suzuki-Takahashi 7 Suzuki-Tanaka 8 (1) It is determined whether or not there is a learning result of “electronic”, “information”, and “Suzuki” in the learning dictionary 5, and if it is determined that there is no learning result, the process proceeds to the next step. (2) There is no learning content related to “tool” related to “electronic”, but there is the following learning content related to “information” and “Suzuki”.

(A) Learning contents related to “information” include “electricity”, “information”, “machine”, and “Sato”: tools. (B) Learning contents related to “Suzuki” include “machine”, "Suzuki": There are tools.

Therefore, the points (a) and (b) will be calculated. In the case of the above (a), "electronic""electricity" 10 "electronic""information" 10 "electronic""machine" 5 "electronic""Sato" 10 "information""electricity" 10 "information""machine" 5 “Information” “Sato” 0 “Suzuki” “Electricity” 0 “Suzuki” “Information” 0 “Suzuki” “Machine” 0 “Suzuki” “Sato” 0, and the evaluation score is 50 points.

In the case of the above (b), “Electronic” “Machine” 5 “Electronic” “Suzuki” 0 “Information” “Machine” 5 “Information” “Suzuki” 10, and the evaluation score is 20 points.

Therefore, the translation of "tool" in this example translates "tool" which is a translation of the technical term dictionary 4 of "information", which is a specialized field having a high evaluation score. In the above embodiment, the user dictionary 3 and the user-related dictionary relation table 9 are used. However, a configuration without the dictionary 3 and the table 9 may be used. In such a machine translation system, operator identification data and environmental data such as a specialized field of "information" are input as a pre-translation operation. Therefore, when a translation word under the operator identification data is selected, the operator identification data And translated words can be learned and stored in the learning dictionary 5, and if there is no learning result in the learning dictionary 5, the technical term dictionary relation table 1
From 0, it is determined whether or not there is a translated word result in the technical term dictionary 4 of, for example, "electricity" or "electronic" which is closely related to "information". When there is a translated word result, these "information", "electricity",
Calculate the relevant scores for each specialized field of “Electronics”, and translate the translated word of the technical term dictionary 4 with the highest evaluation score into the first
It is also possible to select it as a translation.

For the translation of technical terms, the technical term dictionary 4 and the technical term dictionary relation table 10 are mainly used. Therefore, the dictionary 3 and the table 9 are intended for the user, but need not necessarily be limited to the user. The dictionary 3 and the table 9 have a function of an auxiliary dictionary or an auxiliary dictionary necessary for finding a translation of the technical term dictionary 4. I just need.

Further, in the relation tables 9 and 10, for example, the scores of those having strong closeness between the specialized fields are increased, but it is needless to say that the scores of those having strong closeness may be reduced. The point is to specify a terminology dictionary that has a high reputation, not a large or small score.

The present invention is not limited to the above embodiment. For example, each dictionary (2, 3,
4, 5), the control means (21, 22, 23), the relation tables (9, 10), and the functions of the storages (6, 8) are programmed and written in advance on a recording medium such as a CD-ROM. By loading this CD-ROM, the same functions as those of the above embodiment can be realized. The recording medium may be a magnetic tape, a DVD-ROM, a floppy disk, an MO, an MD, a CD-R, a memory card, or the like, other than the CD-ROM.

[0045]

As described above, according to the present invention, the following various effects can be obtained. According to the first and second aspects of the present invention, the translation of a language in a specialty field is evaluated under a predetermined rule while considering a technical term dictionary in a related specialty field related to the specialty field. It is possible to provide a translated word generation method for translating a translated word from one technical term dictionary.

Next, according to the third and fifth aspects of the present invention, the translation result of the same language is translated into the terminology dictionary of the related specialty field closely related to the certain specialty field with respect to the translation of the language of a certain specialty field. If there is, for a certain specialty field and a related specialty field with a translation result, the related points are added from the scores in the above-mentioned technical term dictionary relation table, and a technical term dictionary of one specialty field is specified from the highest evaluation score. Therefore, it is possible to provide a machine translation system capable of translating an appropriate translated word.

Further, according to the third and fifth aspects of the present invention, not only a technical term dictionary of a related specialty field closely related to a specialty field but also a technical term dictionary relation table including, for example, an auxiliary dictionary of a user, etc. Since the evaluation is performed based on the number of relations in the auxiliary dictionary relation table, the translated word is translated from one highly evaluated auxiliary dictionary or one technical term dictionary, so that it is possible to provide a machine translation system capable of translating a more appropriate translated word.

Further, according to the invention of claim 6 to claim 8, the computer reads the program on the recording medium, so that the computer translates a language in a special field from one of the highest evaluation scores. It can identify the specialized terminology dictionary of the field of specialization and translate it appropriately, and can also provide language translation instructions based on the relationship between one auxiliary field and the specialty field, and the relation between other auxiliary fields and the specialty field. Even if a language translation instruction is received below, the translated word can be appropriately translated.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing an embodiment of a machine translation system according to the present invention.

FIG. 2 is a flowchart for explaining the operation of the system shown in FIG. 1;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Translation processing control part 3 ... User dictionary (attached dictionary) 4 ... Technical term dictionary 5 ... Learning dictionary 9 ... User dictionary related table (attached dictionary related table) 10 ... Technical term dictionary related table 21 ... Update control means 22 ... Translated word Generation control means 23 ... Learning control means

Claims (8)

[Claims] Claims: 1. Scores are made in advance according to the closeness between specialized fields, and when receiving a translation instruction in a language of a certain specialized field,
A translated word generation method characterized by adding and evaluating the relevance score for each specialized field including a specialized field closely related to a certain specialized field, and translating a translated term of a highly-evaluated technical term dictionary. 2. The addition evaluation extracts a related technical field having a close relationship with a certain technical field from the score of the technical term dictionary relation table, and, when there is a translation result in the technical term dictionary of the relevant technical field, the evaluation is performed. 2. The translated word generation method according to claim 1, wherein the related points are added and evaluated for each specialty field and the related specialty field, and a technical term dictionary of the specialty field having a high evaluation is specified. 3. A technical term dictionary that defines translations for each specialized field, a technical term dictionary relation table that stores scores in accordance with the closeness of each specialized field, and a translation instruction for a language of a certain specialized field. When received, a related specialty field having a close relationship with a certain specialty field is extracted from the scores of the technical term dictionary relation table, and a related score is added and evaluated for each of the certain specialty field and the related specialty field. A machine translation system comprising: a translated word generation control means for translating a translated word from a technical term dictionary. 4. A technical term dictionary for defining translations of each specialized field, a plurality of auxiliary dictionaries for defining translations of auxiliary fields necessary for translating the translation of the specialized field, A technical term dictionary relation table that is converted into a score according to the relation, and an auxiliary dictionary relation table that stores a score that changes according to the number of uses of the auxiliary field and the special field. When a language translation instruction is received, an update control means for counting the number of times of use of the certain auxiliary field and specialty field and updating the score of the auxiliary dictionary relation table; Means for judging whether a translation instruction of a language based on the language or a translation instruction of a language based on an additional field and a special field which spans another area; In such a case, a related specialty field having a close relationship with the certain specialty field is extracted from the scores in the technical term dictionary relation table, and the related score is added and evaluated for each of the certain specialty field and the related specialty field. If a translation word is translated from the term dictionary and a translation instruction of a language is received under the relation between the other related field and the special field, the related term is obtained from the score stored in the technical term dictionary related table and the related dictionary related table. One additional dictionary and one with high evaluation are evaluated by adding points.
And a translation word generation control means for translating a translation word from one technical term dictionary. 5. A translation word generation control unit extracts a related specialty field having a close relationship with a certain specialty field from the score of the technical term dictionary relation table, and the translation result is stored in the technical term dictionary of the related specialty field. At this time, the related points are added and evaluated for each of the certain specialty field and the related specialty field, a technical term dictionary of one specialty field having the highest evaluation is specified, and a translation word is translated from the technical term dictionary. The machine translation system according to claim 3 or 4. 6. A technical term dictionary for defining translations of each specialty field, means for scoring and storing in a first storage means a score according to the closeness of each specialty field, translation of a language of a certain specialty field When the instruction is received, a related specialty field having a close relationship with a certain specialty field is extracted from the scores of the first storage means, and a related score is added and evaluated for each of the certain specialty field and the related specialty field. And a means for translating a translated word from a technical term dictionary with high quality. 7. A technical term dictionary defining translations in each specialized field, a plurality of auxiliary dictionaries defining translations in an auxiliary field necessary for translating the translations in the specialized field, and a close relationship between the respective specialized fields. Means for storing the score in accordance with the number of times of use of the attached field and the specialty field, second means for storing a score which changes in accordance with the number of uses of the attached field and the specialized field, When receiving a language translation instruction under the association,
Update control means for counting the number of uses of the certain auxiliary field and specialty field and updating the score of the second storage means; and a translation instruction of a language based on one certain one of the auxiliary field and the specialty field, or Means for determining whether a language translation instruction is based on an attached field and a specialty field, and a first storage means when a language translation instruction is received in association with the certain one additional field and a specialty field The relevant specialty field that has a close relationship with the certain specialty field is extracted from the scores of the relevant specialty field, the related score is added and evaluated for each of the certain specialty field and the relevant specialty field, and a translation is translated from one technical term dictionary. When a language translation instruction is received under a relation between an additional field and a special field, the related points are added and evaluated from the points stored in the first storage means and the second storage means, and the evaluation is performed. One with high Accessory dictionary and one computer-readable recording medium recording a program for realizing the translation generation control means for translation of the translation from the terminology dictionary. 8. The translated word generation control means extracts a related specialty field having a close relationship with the certain specialty field from the score of the first storage means, and has a translation result in a technical term dictionary of the related specialty field. At this time, a claim is recorded in which a related point score is added and evaluated for each of the certain specialty field and the related specialty field, a technical term dictionary of one specialty field having the highest evaluation is specified, and a program for translating a translation word from the technical term dictionary is recorded. The computer-readable recording medium according to claim 6 or 7.