JP2007264858A - Personal name sexuality determination program, machine translation program, personal name sexuality determination device, machine translation device, personal name sexuality determination processing method and machine translation processing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a new technique that can select appropriate pronouns required in machine translation by enabling determination of sexuality of personal names consisting only of family names, and various personal names included in sentences. <P>SOLUTION: Sexuality of personal names are basically determined in view of context information about the personal names. Further, sexuality of personal names are determined in consideration of the notation of the personal names. Sexuality of personal names are also in consideration of the sexualization of similar personal names included in a set of sentences where the same personal name may appear. The processing can determine sexuality even for personal names that cannot be uniquely determine sexuality, and can determine sexuality even personal names consisting only of family names. The sexuality determination can be implemented without a comprehensive personal name sex dictionary. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a personality gender determination device that determines the gender of a personal name included in a sentence to be processed, a processing method thereof, a personality gender determination program used for realizing the personality gender determination device, and a personality gender determination technique thereof. The present invention relates to a machine translation apparatus that performs machine translation using the machine translation method, a processing method thereof, and a machine translation program used to realize the machine translation apparatus.

  In machine translation, it is important to select an appropriate pronoun. In particular, in Japanese-English machine translation, it is necessary to supplement pronouns that are not found in Japanese or to convert pronouns that are described with proper names (person names) into pronouns.

  To that end, gender information for each person's name is required, but in the current machine translation, gender information is given only for a very limited number of typical names. Giving has not been realized.

  In Patent Document 1 shown below, when performing machine translation, the proper nouns and common nouns related to people in the translated sentence are searched based on the information in the dictionary part to identify the gender of those nouns. However, because the information in the dictionary is limited and cannot cover proper nouns and common nouns related to people, it is not actually possible to provide gender information for a wide range of personal names. Is the actual situation.

  From now on, according to the prior art, it is not possible to select an appropriate pronoun when executing machine translation.

In addition, as a technique related to the present invention, Patent Document 2 shown below describes an invention in which a proper noun expressed in alphabet is converted into a corresponding proper Chinese kanji proper noun using a neural network. In this invention, when the proper noun of the alphabet notation has a flag indicating a female name such as Miss, it is suitable for a woman according to a unit responding to the female name provided in the input layer of the neural network. It adopts the structure of converting to proper nouns of Chinese kanji.
Japanese Patent Laid-Open No. 62-269264 JP-A-5-324718

  In order to select an appropriate pronoun in Japanese-English machine translation, the gender information of each person's name in Japanese is required, and as a solution, a dictionary for person name and gender is prepared as described in Patent Document 1. It is conceivable to use a method.

  However, using the method of determining the gender of a person name by preparing a personality gender dictionary cannot actually determine the gender of the person name.

  That is, it is practically difficult to create a dictionary that comprehensively describes all names including foreign names. In addition, there are names that cannot uniquely determine the gender, and because the name of the surname alone cannot determine the gender at all, relying solely on the personal name gender dictionary actually determines the gender of the name It cannot be realized.

  The present invention has been made in view of such circumstances, and it is possible to determine the gender of a personal name with only a surname or a wide range of personal names included in a sentence, and an appropriate pronoun required for machine translation. The purpose is to provide a new technology that makes it possible to realize this selection.

[1] Configuration of personality gender determination apparatus of the present invention The personality gender determination apparatus of the present invention that implements the personality gender determination program of the present invention performs a process of determining the gender of a personal name included in a sentence to be processed. In order to enable selection of appropriate pronouns required for machine translation, (a) extracting context information around the person name to be judged, or surrounding the person name to be judged (B) a determination means for determining the gender of a person name to be determined based on the information extracted by the extraction means, and (c) a processing target. Comparing the information on the gender determined by the determination unit for similar names included in the sentence, and a correction unit for correcting the determined gender information based on the comparison result.

  Each of the above processing means can also be realized by a computer program. This computer program is provided by being recorded on an appropriate computer-readable recording medium or provided via a network, and is used when implementing the present invention. The present invention is realized by being installed and operating on a control means such as a CPU.

  The personality gender determination apparatus of the present invention configured as described above is based on determining the gender of the personal name using context information around the personal name included in the sentence to be processed. In addition, the gender of the person's name is determined in consideration of the notation of the person's name, or the gender of the person's name is determined in consideration of gender information about similar person names included in the sentence to be processed. .

[2] Configuration of the machine translation apparatus of the present invention The machine translation apparatus of the present invention that implements the machine translation program of the present invention can realize selection of an appropriate pronoun necessary for machine translation. (A) Using the gender of a person's name contained in a sentence subject to machine translation as a processing target, extract the context information around that person's name, or extract the context information around that person's name and the notation of that person's name Extraction means, (b) determination means for determining the gender of a person name included in a sentence to be machine-translated based on information extracted by the extraction means, and (c) included in a sentence to be machine-translated For similar person names, the sex information determined by the determining means is compared, and based on the comparison result, the correcting means for correcting the determined sex information, and (d) the sex determined by the determining means Information When the correct means corrects the sex information, the correct means includes selection means for selecting pronouns necessary for machine translation based on the corrected sex information.

  Each of the above processing means can also be realized by a computer program. This computer program is provided by being recorded on an appropriate computer-readable recording medium or provided via a network, and is used when implementing the present invention. The present invention is realized by being installed and operating on a control means such as a CPU.

  The machine translation apparatus of the present invention configured as described above is based on determining the gender of the person name using context information around the person name included in the sentence to be machine translated. In addition, the gender of the person's name is determined in consideration of the notation of the person's name, or the gender of the person's name is determined in consideration of the gender information about the similar person name included in the sentences subject to machine translation. Then, based on the determined gender, processing is performed so that pronouns necessary for machine translation are selected.

  In the present invention, based on determining the gender of the person name using the context information around the person name, in addition to this, the gender of the person name can be determined in consideration of the notation of the person name. , The gender of the person name is determined in consideration of the gender judgment result of similar person names included in a single sentence where the same person name may appear. Gender can be determined even for names that cannot be determined, gender can also be determined for names with only surnames, and gender can be determined without preparing a comprehensive personality gender dictionary. become.

  Thus, according to the present invention, it is possible to realize selection of appropriate pronouns required for machine translation for a wide range of personal names without preparing comprehensive personal gender information.

  Hereinafter, the present invention will be described in detail according to embodiments.

  FIG. 1 illustrates an embodiment of a machine translation apparatus 1 including the present invention.

  As shown in this figure, the machine translation apparatus 1 of the present invention recognizes an input unit 10 for inputting a sentence to be machine-translated from a file or a keyboard, and a personal name part included in the sentence input by the input unit 10. In addition, the proper noun recognition unit 11 that outputs the personal name portion with a tag or the like, the personal gender determination unit 12 that determines the gender of the personal name recognized by the proper noun recognition unit 11, and the determination of the personal gender determination unit 12 Based on the result, a machine translation unit 13 that machine translates a sentence input by the input unit 10 and an output unit 14 that outputs a translation result of the machine translation unit 13 are provided.

  Here, for the proper noun recognition unit 11, a conventional proper noun recognition system can be used as it is.

  The personality gender determination unit 12 performs a process of determining the gender of the personal name recognized by the proper noun recognition unit 11, a determination rule storage unit 120, a personality gender dictionary 121, a preprocessing unit 122, and a gender determination unit. 123, a gender determination unit 124, and a gender information adding unit 125.

  The determination rule storage unit 120 describes a relationship that is established between the context information around the name of the person and the gender of the person name, and when there is such context information, the gender of the person name is as follows. At the same time, a determination rule that describes the certainty of the relationship is stored. This determination rule is composed of, for example, a decision tree having context information around a person name as a node and the gender of the person name as a leaf.

  FIG. 2 illustrates an example of a determination rule composed of such a decision tree.

  In this figure, if the notation of the morpheme before the name of the person is “noun“ sister ”” and the notation of the morpheme before the name of the person is “no” of the particle, the gender of the name is For women, the degree of certainty is “1.0”, or “the morpheme in front of the person ’s name is“ Father ”” and the morpheme in front of the person ’s name is “ If it is “・” ”, the gender of the person's name is male and the certainty is“ 1.0 ”” or “the morpheme notation after the person ’s name is“ Noun ”in the particle, If the morpheme notation after the name of the person is “noun“ wife ””, the gender of the name is male and the certainty is “1.0” as an example. ing.

  The personal name gender dictionary 121 describes gender information of the personal name in association with the notation of the personal name and manages dictionary data describing the certainty of the relationship.

  FIG. 3 shows an example of dictionary data managed by the personality gender dictionary 121.

  In this figure, “If you have a personal name notation of“ Taro ”, the gender of the personal name is male and the certainty is“ 1.0 ”” or “If you have a personal name notation of“ Taro ” If the person ’s gender is male and the confidence is “0.8”, or “the name is“ Akiko, ”the person ’s gender is female and the confidence is“ For example, dictionary data such as “1.0” ”or“ if the name of the child is “child” has a gender of female name and the certainty is “0.5”. ” Show.

  Here, in the machine translation apparatus 1 according to the present invention, the determination process of the personal name is performed using the determination rule stored in the determination rule storage unit 120. It is not necessary to prepare dictionary data comprehensively.

  The preprocessing unit 122 divides the text input by the input unit 10 tagged with the personal name into articles as preprocessing for gender determination of personal names, and executes morphological analysis on the texts of those articles.

  In this article dividing process, a process of dividing a sentence input by the input unit 10 into an article that is a single sentence unit (a single sentence unit in which the same person name may appear). Divided into articles according to a pre-determined article format (for example, one article has a title), divided into articles according to instructions from the user, etc. Use to divide into articles.

  In this morpheme analysis process, a sentence divided into articles is divided into morphemes that are the units of the preceding and following contexts, and information such as part of speech of each divided morpheme is analyzed. Here, for this morpheme analysis processing, a conventional morpheme analysis system can be used as it is.

  The gender determination unit 123 sets the person name recognized by the proper noun recognition unit 11 described in the article divided by the preprocessing unit 122 as a determination target, the determination rule stored in the determination rule storage unit 120, and the personal name gender dictionary 121. The gender of the person name is determined based on the context information around the person name and the notation of the person name while referring to the dictionary data to be managed.

  The gender determination unit 124 compares the determination result of the gender determination unit 123 for each person name described in the article with the determination result of the gender of a similar person name described in the article, based on the result Then, the determination result of the gender determination unit 123 is fixed as it is, or the gender determined by the gender determination unit 123 is fixed in reverse.

  The gender information adding unit 125 adds the gender information determined by the gender determining unit 124 to each person name described in the article. As a method for this assignment, the tag assigned by the proper noun recognizing unit 11 is rewritten with a tag representing gender.

  In the personality gender determination unit 12 configured as described above, the sentence input by the input unit 10 is divided into articles, and for each person name recognized by the proper noun recognition unit 11 in units of the article, Based on the context information and the notation of the person's name, a process of determining the gender of the person's name is performed.

  In response to the determination result of the gender of the personal name, the machine translation unit 13 selects the appropriate pronoun by using the gender information of the personal name given by the personal name gender determination unit 12 and reads the sentence input by the input unit 10. Machine translation.

  In traditional Japanese-English machine translation, gender information is not given to proper nouns, so when translating the sentence "Akiko shook his right hand", "Akiko shook his right hand" Interpret it to meaning "Akiko shook one's right hand", but because you cannot select the pronoun of "one's", "Akiko shook my right hand", "Akiko shook the right hand", "Akiko shook his" There was the inconvenience of translating to "right hand".

  In contrast, in the present invention, since the proper noun “Akiko” is given gender information of female, the machine translation unit 13 does not use “my” or “his” in this case, but the exact “her”. The pronoun "" can be selected, so that it can be accurately translated as "Akiko shook her right hand".

  FIG. 4 illustrates an example of a processing flow executed by the personality gender determination unit 12.

  Next, according to this processing flow, processing executed by the personality gender determination unit 12 will be described in detail.

  When the proper name noun recognizing unit 11 recognizes the personal name part included in the sentence input by the input unit 10 and outputs the personal name part with a tag or the like added thereto, the personal name gender determining unit 12 displays As shown, first, in step S10, the input document is divided into articles.

  Subsequently, in step S11, one unprocessed article is selected, and in subsequent step S12, morphological analysis is performed on the sentence of the selected article.

  Subsequently, in step S13, one unprocessed person name is selected from the person names described in the selected article.

  Subsequently, in step S14, context information around the selected person name is specified, and a determination rule that matches the specified context information is extracted from the determination rules stored in the determination rule storage unit 120. On the basis of the certainty defined in the determination rule, the certainty for determining male and the certainty for determining female are obtained for each gender.

  At this time, when extracting a plurality of determination rules, by summing the certainty levels defined in those determination rules, for each gender, the certainty level for determining male and the female level for determining female Try to get confidence and.

  Subsequently, in step S15, by referring to the dictionary data managed by the personal name gender dictionary 121, dictionary data matching the notation of the selected personal name is extracted, and the certainty factor defined by the extracted dictionary data is obtained. Based on each gender, the certainty factor in the case of determining that it is male and the certainty factor in the case of determining that it is female are obtained.

  At this time, in the case of extracting a plurality of dictionary data, by summing the certainty defined in those dictionary data, for each gender, the certainty for determining male and female Try to get confidence and.

  Subsequently, in step S16, for each gender, the certainty factor of the determination result obtained in step S14 and the certainty factor of the determination result obtained in step S15 are summed, and the gender having the higher total value is selected. Judged as the gender of the name

  For example, when determining the gender of the personal name “Yamada Taro” possessed by the sentence “Father Taro Yamada became prime minister”, as shown in FIG. 5, in step S14, as shown in FIG. According to the determination rule, the certainty factor that “Taro Yamada” is male is 1.0 and the certainty factor that is female is 0, and in step S15, the personality gender dictionary 121 as shown in FIG. According to the dictionary data, the certainty factor that “Taro Yamada” is male is 1.8 and the certainty factor that is female is 0. In step S16, the total certainty factor that “Taro Yamada” is male. Is 2.8, and based on the determination result that the total value of the certainty of being a woman is 0, “Taro Yamada” is finally determined to be a man.

  Subsequently, in step S17, the determination result obtained in step S16 is stored in the comparison buffer. In subsequent step S18, it is determined whether or not the genders of all personal names described in the selected article have been determined. When it is determined that all the names have not been determined, the process returns to step S13.

  In this way, when the processes of step S13 to step S18 are repeated, when it is determined in step S18 that the genders of all personal names described in the selected article have been determined, the process proceeds to step S19. Extracting partial matches for the names stored in the comparison buffer and collecting the gender determination results for those names, for each name listed in the selected article, gender of similar names Collect judgment results.

  Subsequently, in step S20, the gender of each personal name described in the selected article is determined by performing a majority decision on the collected gender determination results.

  For example, as shown in FIG. 6, there are four personal names “Yamada” and one personal name “Taro Yamada” in the same article. "Yamada" is determined to be male, the second "Yamada" is determined to be female, "Taro Yamada" is determined to be male, the third "Yamada" is determined to be male, and the fourth When “Yamada” is determined to be male, the majority of these gender determination results are determined to determine that “Yamada” and “Taro Yamada” are male.

  Here, if there is no similar person name, the gender determined in step S16 is determined as it is without making a majority vote.

  Subsequently, in step S21, the gender information determined in step S20 is assigned to each person name described in the selected article, and in step S22, the text of the selected article is stored in the output buffer.

  Subsequently, in step S23, the comparison buffer is cleared in order to start processing for the next article. In subsequent step S24, it is determined whether or not all the articles divided in step S10 have been determined. When it is determined that there is an undetermined article, the process returns to step S11.

  On the other hand, when it is determined that all the articles divided in step S10 have been determined according to the determination process in step S24, the process proceeds to step S25, and the contents of the output buffer are output to the machine translation unit 13.

  In this way, the personality gender determination unit 12 divides the sentence input by the input unit 10 into articles, and for each personal name recognized by the proper noun recognition unit 11 in units of the article, the context around the personal name Processing is performed so as to determine the gender of the person name based on the information and the notation of the person name.

  In response to the determination result of the gender of the personal name, the machine translation unit 13 selects the appropriate pronoun by using the gender information of the personal name given by the personal name gender determination unit 12 and reads the sentence input by the input unit 10. Machine translation can be realized.

  In the exemplary embodiment illustrated in FIG. 1, the gender determination unit 123 performs processing to determine the gender of the personal name using the determination rule stored in the determination rule storage unit 120 and the dictionary data managed by the personal name gender dictionary 121. However, it is also possible to determine the gender of the name using other methods. For example, a classifier method based on machine learning can be used.

When determining the gender of a person's name using this classifier method by machine learning, the gender determination unit 123 has a machine learning function, and at the time of learning, for example,
{Indication of the morpheme before the name of the person, part of speech of the morpheme before the name of the person,
The notation of the morpheme before the name of the person, the part of speech of the morpheme before the name of the person,
Notation of person name, notation of first morpheme of person name, notation of last morpheme of person name,
The notation of the morpheme after the name, the part of speech of the morpheme after the name,
Notation of the morpheme after the name of the person, part of speech of the morpheme after the name of the person,
Gender of the name}
Is stored.

For example, based on the sentence “Father Taro Yamada became Prime Minister” and gender information that “Taro Yamada” is a male,
{Father, noun, .., symbol, Taro Yamada, Taro Yamada, particle, prime minister, noun, male}
Is generated and stored.

  When the gender determination unit 123 accumulates the vector data, based on the accumulated vector data, for example, “if the notation of the previous morpheme is“ father ”” or “if the last notation of the personal name is“ Taro ” A decision tree is generated by generating a node such as “If”, and generating a tree having a leaf as the gender of the name in association with the node.

  Then, after generating the decision tree, the gender determination unit 123 determines the gender of the person name based on the vector data about the person name to be determined for gender and the generated decision tree when entering the actual determination process. To do.

  This classifier generation algorithm by machine learning has been introduced in various documents such as “Data analysis by AI, translated by JR Kinran, Koichi Furukawa, pp 4-8, 1995”. By using vector data in such a data format, it is possible to determine the sex of a person's name.

  Next, an experiment conducted for verifying the effectiveness of the present invention will be described.

  This experiment was performed by determining the gender of a person's name using the classifier method by machine learning. As the data format of the vector data used for the generation of the decision tree, the one described above is used. Note that the majority vote method described above is not used because it was performed to check the accuracy rate.

  For experimental data, use 7 days (about 8,000 sentences) of newspaper articles labeled with names and gender as learning data, and use 1 day of newspaper articles (about 1,200 sentences) as test data for sex determination. I went there. FIG. 7 illustrates an example of experimental data.

  In this experiment, 403 person names were subject to gender determination, and the sexes of 367 person names showed correct answers, so the correct rate was 91% (= 367/403). Of these 403 person names, 84% are men, and if all are determined to be men without doing anything, the accuracy is 84%. Therefore, it was verified that the effects of the present invention were effective.

  Further, in this experiment, 172 people with only surnames were judged, and the sex of 153 people showed correct answers, so the correct rate was 89% (= 153/172). . Of these 153 surnames, 83% are males, and if all are determined to be males without doing anything, the accuracy is 83%. Thus, the effect of the present invention was verified. From this, it can be verified that the present invention is effective for the gender determination of the personal name of only the surname that cannot be determined only by the personal name gender dictionary 121.

  In this experiment, the above-described majority method is not used, so the accuracy rate is further increased by using the above-described majority method.

  In the present invention, the confidence rule is defined for the judgment rule stored in the judgment rule storage unit 120 and the dictionary data managed by the personality gender dictionary 121. It is preferable to learn to improve the result.

  As a method of learning the certainty factor, prepare data labeled with a person's name and gender, 90% of the data is learned, the remaining 10% of the data is used for testing, and gender determination is performed. It is possible to use a method of learning by obtaining a certainty factor that allows the best judgment.

  When such a method is used, a total of 10 sex determination experiments can be performed by changing 10% of the test data to another one in order, so the confidence that the result is the best is learned. By doing so, it becomes possible to learn an appropriate certainty level.

  According to the present invention, when it is necessary to determine the gender of a person name included in a sentence, it becomes possible to determine the gender of only a surname or a wide range of person names. It can be applied to various fields such as machine translation that needs to be determined.

  (Supplementary note 1) A personality gender determination program for performing a process for determining the gender of a personal name included in a sentence to be processed, the computer extracting the context information around the personal name, A personal name gender determination program for functioning as a determination unit for determining gender of the personal name based on information extracted by an extraction unit.

  (Supplementary note 2) A personality gender determination program for performing a process for determining the gender of a personal name included in a sentence to be processed, the computer extracting context information around the personal name and a description of the personal name A personal sex determination program for functioning as an extraction means and a determination means for determining the sex of the personal name based on the information extracted by the extraction means.

  (Supplementary note 3) In the personal name sex determination program according to supplementary note 2, the determination means includes a certainty factor of the gender determination result performed based on the context information extracted by the extraction means, and a personal name expression extracted by the extraction means The personal name gender determination program characterized by calculating a total value with the certainty of the gender determination result made based on the above and determining the gender of the personal name based on the total value.

  (Supplementary Note 4) In the personal name gender determination program according to any one of supplementary notes 1 to 3, the computer further includes information on the gender determined by the determination means for a similar personal name included in the sentence to be processed. And a personality gender determination program characterized in that, based on the comparison result, the information on the determined gender is corrected.

  (Supplementary note 5) A machine translation program that performs processing for machine translation of a sentence, in which context information around the person's name is processed using the gender of the person name included in the sentence to be machine translated Extraction means for extracting, based on the information extracted by the extraction means, determination means for determining the gender of the personal name, and based on the gender determined by the determination means, pronouns required for machine translation A machine translation program for functioning as a selection means for performing selection.

(Supplementary note 6) A machine translation program that performs processing for machine translation of a sentence, and uses the computer as a processing target for the gender of a person name included in the sentence to be machine translated, and context information around the person name And an extracting means for extracting the notation of the person name,
Determination means for determining the gender of the person name based on the information extracted by the extraction means;
A machine translation program for functioning as a selection means for selecting a pronoun necessary for machine translation based on the gender determined by the determination means.

  (Supplementary note 7) In the machine translation program according to supplementary note 6, the determination means includes a certainty factor of the gender determination result performed based on the context information extracted by the extraction means and a personal name notation extracted by the extraction means. A machine translation program characterized by calculating a total value with the certainty of the determination result of gender performed based on the result and determining the gender of the personal name based on the total value.

  (Supplementary note 8) In the machine translation program according to any one of supplementary notes 5 to 7, the computer and the gender determined by the determination means for a similar person name included in a sentence to be machine-translated. A machine translation program characterized by comparing information and functioning as correcting means for correcting the determined sex information based on the comparison result.

  (Supplementary note 9) A personality gender determination apparatus for determining the gender of a personal name included in a sentence to be processed, the extraction means for extracting context information around the personal name and the representation of the personal name, A personality gender determination device comprising: a determination unit that determines the gender of the personal name based on the extracted information.

  (Supplementary Note 10) A machine translation device for machine-translating a sentence, and extracting context information and the notation of the person's name around the person's name by using the gender of the person's name included in the sentence to be machine-translated Based on information extracted by the extraction means, extraction means, determination means for determining the gender of the personal name, and selection of pronouns required for machine translation based on the gender determined by the determination means A machine translation device comprising a selection unit.

  (Supplementary Note 11) A personality gender determination processing method executed by a personality gender determination apparatus for determining the gender of a personal name included in a sentence to be processed, wherein context information around the personal name and a description of the personal name are extracted And determining the gender of the personal name based on the extracted information.

  (Additional remark 12) It is the machine translation processing method performed with the machine translation apparatus which machine-translates a text | text, Comprising: The context information in the circumference | surroundings of the person name made into the object of the gender of the person name contained in the sentence used as the machine translation object Extracting the notation of the person's name, determining the gender of the person's name based on the extracted information, and selecting pronouns required for machine translation based on the determined gender, A machine translation processing method.

1 is an embodiment of a machine translation device including the present invention. It is a figure which shows an example of the determination rule preserve | saved at the determination rule memory | storage part. It is a figure which shows an example of the dictionary data managed by a personal name sex dictionary. It is a processing flow which a person name gender determination part performs. It is explanatory drawing of the process which a person name sex determination part performs. It is explanatory drawing of the process which a person name sex determination part performs. It is a figure which shows an example of the experimental data performed for verifying the effectiveness of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Machine translation apparatus 10 Input part 11 Proper noun recognition part 12 Personality gender determination part 13 Machine translation part 14 Output part 120 Judgment rule memory | storage part 121 Personality sex dictionary 122 Pre-processing part 123 Gender judgment part 124 Gender determination part 125 Gender information provision part

Claims (10)

A personal name gender determination program for performing a process for determining the gender of a personal name included in a sentence to be processed,
Computer
Extracting means for extracting context information around the person name;
A personality gender determination program for functioning as a determination unit for determining gender of the personal name based on information extracted by the extraction unit. A personal name gender determination program for performing a process for determining the gender of a personal name included in a sentence to be processed,
Computer
Extraction means for extracting contextual information around the person name and a description of the person name;
A personality gender determination program for functioning as a determination unit for determining gender of the personal name based on information extracted by the extraction unit. In the personal gender determination program according to claim 1 or 2,
The computer further compares the sex information determined by the determination means for similar names included in the sentence to be processed, and corrects the determined sex information based on the comparison result. To function as a correction means,
A personal name gender determination program. A machine translation program that performs processing for machine translation of a sentence,
Computer
An extraction means for extracting the context information around the person name, with the gender of the person name included in the sentence to be machine-translated;
Determination means for determining the gender of the person name based on the information extracted by the extraction means;
A machine translation program for functioning as a selection means for selecting a pronoun necessary for machine translation based on the gender determined by the determination means. A machine translation program that performs processing for machine translation of a sentence,
Computer
An extraction means for extracting the context information around the person name and the notation of the person name, with the gender of the person name included in the sentence to be machine-translated as a processing target;
Determination means for determining the gender of the person name based on the information extracted by the extraction means;
A machine translation program for functioning as a selection means for selecting a pronoun necessary for machine translation based on the gender determined by the determination means. In the machine translation program according to claim 4 or 5,
The computer further compares the sex information determined by the determination means with respect to similar person names included in the sentences to be machine-translated, and the determined sex information is obtained based on the comparison result. To function as a correction means to correct,
Features machine translation program. A personality gender determination device that determines the gender of a personal name included in a sentence to be processed,
Extraction means for extracting contextual information around the person name and a description of the person name;
A determination unit that determines gender of the person name based on the information extracted by the extraction unit;
Characteristic gender determination device. A machine translation device that machine translates sentences,
An extraction means for extracting the context information around the person name and the notation of the person name, with the gender of the person name included in the sentence to be machine-translated as a processing target;
Determination means for determining the gender of the person name based on the information extracted by the extraction means;
A selection means for selecting a pronoun necessary for machine translation based on the gender determined by the determination means,
A machine translation device. A personality gender determination processing method executed by a personality gender determination device that determines the gender of a personal name included in a sentence to be processed,
Extracting contextual information around the person name and the notation of the person name,
Determining the gender of the person name based on the extracted information;
Characteristic name gender determination processing method. A machine translation processing method executed by a machine translation device that machine translates sentences,
Using the gender of the name of the person included in the text subject to machine translation as the processing target, the context information around the name and the notation of the name are extracted
Based on the extracted information, determine the gender of the person name,
Based on the determined gender, selecting pronouns necessary for machine translation,
A machine translation processing method.

Images