JP3251032B2 - Japanese analyzer - Google Patents
Japanese analyzerInfo
- Publication number
- JP3251032B2 JP3251032B2 JP17621591A JP17621591A JP3251032B2 JP 3251032 B2 JP3251032 B2 JP 3251032B2 JP 17621591 A JP17621591 A JP 17621591A JP 17621591 A JP17621591 A JP 17621591A JP 3251032 B2 JP3251032 B2 JP 3251032B2
- Authority
- JP
- Japan
- Prior art keywords
- noun
- dependency
- range
- unit
- sentence
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Machine Translation (AREA)
- Document Processing Apparatus (AREA)
Description
【０００１】[0001]
【産業上の利用分野】本発明は、日本語文の係り受け構
造を機械的に解析する日本語解析装置に関し、特に複数
の名詞を含む名詞句内のいずれかの名詞に係る埋込み文
を有する日本語文の係り受け構造を解析する日本語解析
装置に関する。BACKGROUND OF THE INVENTION This invention relates to Japanese analyzer for mechanically analyzing the dependency structure of the Japanese sentence, Japanese especially those with embedded statements according to any of the nouns in noun phrase that includes a plurality of nouns Japanese Analysis for Dependency Structure of Japanese Sentences
Related to the device .
【０００２】[0002]
【従来の技術】従来、日本語文の係り受け構造を機械的
に解析する場合、まず日本語文を形態素に分解し、日本
語辞書等のデ−タベ−スを参照して各形態素にその構文
情報や意味情報を付加する形態素解析処理を行う。そし
て、形態素に付加された構文情報や意味情報を利用し、
所定の規則に従って句構造や係り受け関係を解析する構
文解析処理を行う。なお、以下では、日本語文のより正
確な意味構造を得る必要性、及び構文解析の制御にアル
ゴリズムを簡単にできることに鑑み、係り受け関係の解
析を文節ごとではなく形態素ごとに行うこととして構文
解析処理についての説明をする。2. Description of the Related Art Conventionally, when mechanically analyzing the dependency structure of a Japanese sentence, the Japanese sentence is first decomposed into morphemes, and each morpheme is provided with syntax information by referring to a database such as a Japanese dictionary. And morphological analysis processing for adding semantic information. Then, using the syntax information and semantic information added to the morpheme,
A syntax analysis process for analyzing a phrase structure and a dependency relationship according to a predetermined rule is performed. In the following, considering the necessity of obtaining a more accurate semantic structure of a Japanese sentence and the fact that an algorithm can be easily controlled for syntactic analysis, the dependency analysis is performed for each morpheme, not for each clause. The processing will be described.
【０００３】ところで、係り受け構造を解析するべき日
本語文が、埋込み文が複数の名詞を含む名詞句内のいず
れかの名詞に係りかつ該埋込み文が係る可能性のある名
詞が複数あるといった構造を有する場合、上記従来の構
文解析処理では埋込み文が名詞句中のどの名詞に係るか
について正確な係り受け関係の解析をすることは困難で
あった。A Japanese sentence whose dependency structure is to be analyzed has a structure in which an embedded sentence relates to any one of the noun phrases in a noun phrase including a plurality of nouns, and there is a plurality of nouns to which the embedded sentence may relate. In the conventional parsing process described above, it is difficult to accurately analyze the dependency relationship as to which noun in the noun phrase the embedded sentence relates to.
【０００４】そのため、従来の構文解析処理において
は、上述のような構造を有する日本語文に対して、上記
埋込み文が係り得る最初の名詞にかけてしまったり、上
記埋込み文が係り得る全ての名詞にそれぞれかけた場合
の各解析結果を並行的に出力しておき、最終的に人間が
正しい係り受け関係を判断したり、修正したりしてい
た。また、解析結果の精度を上げるため、デ−タ−ベ−
スに事例を蓄積し、埋込み文が係り得る名詞が出てくる
度に上記事例を参照して係り受け関係の正当性を判断
し、必要に応じて埋込み文の係り先を修正するといった
処理をしていた。[0004] Therefore, in the conventional parsing processing, a Japanese sentence having the above-mentioned structure is applied to the first noun to which the embedded sentence can relate, or to all the nouns to which the embedded sentence can relate. Each analysis result in the case of multiplication is output in parallel, and finally a human judges or corrects a correct dependency relationship. In order to improve the accuracy of the analysis results,
Each time a noun to which an embedded sentence can relate appears, the process refers to the above case to determine the validity of the dependency relationship, and if necessary, corrects the modification destination of the embedded sentence. Was.
【０００５】[0005]
【発明が解決しようとする課題】上述したように、従来
の構文解析処理では、埋込み文が複数の名詞を含む名詞
句内のいずれかの名詞に係りかつ該埋込み文が係る可能
性のある名詞が複数あるといった構造を有する日本語文
に対して正確な係り受け関係の解析ができず、人間の判
断や修正を必要とするという欠点があった。As described above, in the conventional parsing process, an embedding sentence relates to any one of the noun phrases in a noun phrase including a plurality of nouns, and the noun that the embedding sentence may be related to. There is a drawback in that it is not possible to accurately analyze dependency relations for a Japanese sentence having a structure such as a plurality of sentences, and human judgment or correction is required.
【０００６】また、デ−タベ−スに事例を蓄積し、これ
を参照して係り受け関係の正当性を判断する場合にも、
正当な係り先と判断される名詞が複数出てきたときに、
それらのうちのいずれが最も正しい係り先かということ
を決めるには、従来のように埋込み文が係る可能性のあ
る名詞が出てくる度に上記事例を参照して係り受け関係
の正当性を判断していたのでは不十分であり、選択の対
象となっている埋込み文が係る可能性のある複数の名詞
の中から一つを特定する手順が必要になる。本発明は、
上記従来の欠点を解決し、上記のような日本語文に対す
る解析を正確かつ効率的に行う日本語解析装置を提供す
ることを目的とする。[0006] Further, when accumulating cases in a database and referring to this to judge the validity of the dependency relationship,
When multiple nouns appear to be legitimate,
To determine which of them is the most correct dependency, refer to the above case and check the validity of the dependency relationship every time a noun that may be involved in the embedded sentence appears as in the past. Judgment is not sufficient, and a procedure for specifying one from a plurality of nouns that may be involved in the embedded sentence to be selected is required. The present invention
An object of the present invention is to solve the above-mentioned conventional drawbacks and to provide a Japanese-language analysis apparatus for accurately and efficiently analyzing Japanese sentences as described above.
【０００７】[0007]
【課題を解決するための手段】上記の目的を達成するた
め、本発明は、入力した日本語文に対し形態素解析処理
を行なう形態素解析部と、上記形態素解析処理のなされ
た日本語文の形態素の係り受け関係を判断し、形態素の
係り受け関係を表わす解析木を作成する構文解析部と、
上記構文解析部が参照するデータベースとして、係り受
け関係の可否に関する判定法や係り方のルールが格納さ
れた構文解析ルール集、形態素によって名詞句の範囲が
終了したか否かを判定するルールが格納された名詞句範
囲判定ルール集、及び埋込み文の種類や名詞句の範囲内
の名詞の種類、名詞間の関係等に基づいて埋込み文の述
語が名詞句の範囲内のいずれの名詞に係るかを規定した
ルールが格納された係り先決定ルール集とを備えてなる
日本語解析装置において、上記構文解析部は、形態素間
に係り受け関係が成立するか否かを係り受け関係の可否
に関する判定法や係り方のルールが格納された構文解析
ルール集に従って判定する係り受け判定部と、該係り受
け判定部の判断に基づいて形態素の係り受け関係を表わ
す解析木を作成する解析木作成部とを備えると共に、上
記係り受け判定部が上記解析木作成部の作成した解析木
によって表わされる埋込み文と該埋込み文の後方の名詞
との係り受け関係を判定した際に、該名詞までで上記埋
込み文が係る可能性のある名詞の存在する範囲が終了し
たか否かを名詞句範囲判定ルール集に従って判定する名
詞句範囲判定部と、該名詞句範囲判定部が上記埋込み文
が係る可能性のある名詞の範囲が終了したと判定した場
合に上記埋込み文が上記名詞の範囲内のいずれの名詞に
係るかを係り先決定ルール集に従って決定する係り先決
定部とを備えてなり、上記解析木作成部は、上記名詞句
範囲判定部の判定に応じて、名詞に上記埋込み文が係る
可能性がある旨の情報及び上記埋込み文と名詞との係り
の属性に関する情報を上記解析木に付加し、または上記
係り先決定部の判断及び上記解析木に付加した上記係り
の属性に関する情報に基づき解析木を作成することを特
徴とする。In order to achieve the above object, the present invention relates to a morphological analysis unit for performing a morphological analysis process on an input Japanese sentence, and a morphological analysis of the Japanese sentence subjected to the morphological analysis process. A parsing unit that determines a dependency relationship and creates a parse tree representing a morpheme dependency relationship;
As the database referred to by the parsing unit, a collection of parsing rules in which a determination method regarding whether or not a dependency relationship is possible and rules of a dependency method are stored, and a rule for determining whether or not the range of a noun phrase is terminated by a morpheme are stored. The set of noun phrase range determination rules, the type of the embedded sentence, the type of the noun in the range of the noun phrase, the relation between the nouns, etc., which predicate of the embedded sentence relates to which noun in the range of the noun phrase in Japanese analyzing device including a dependency destination determination rule collection specified rule is stored, the syntax analysis unit, the determination regarding whether the received dependency whether received relates between morphemes relationship is established relationship A dependency determining unit that determines according to a syntax analysis rule collection in which rules of law and dependency are stored, and a parse tree representing a morpheme dependency based on the determination of the dependency determining unit. A parse tree creating unit, and when the dependency determining unit determines a dependency relationship between an embedded sentence represented by the parse tree created by the parse tree creating unit and a noun behind the embedded sentence, A noun / phrase range determining unit for determining whether or not the range of the noun to which the embedded sentence may be related up to the noun according to a collection of noun / phrase range determining rules; and A determination unit that determines which of the nouns in the range of the noun the embedded sentence pertains to in accordance with a collection of determination rules, when it is determined that the range of the noun that may be related has ended. The parse tree creating unit, according to the determination of the noun phrase range determination unit, the information that there is a possibility that the embedded sentence is related to the noun, and the information about the attribute of the relationship between the embedded sentence and the noun, Parse tree Pressurized to, or characterized by creating a parse tree based on information relating to the dependency of the attribute judgment and added to the analysis tree of the dependency destination determining unit.
【０００８】請求項２の発明は、請求項１の発明におい
て解析木作成部が、係り受け判定部が埋込み文との係り
受け関係を判定した名詞までで上記埋込み文が係る可能
性のある名詞の存在する範囲が終了しないと名詞句範囲
判定部が判定した場合に、上記埋込み文と名詞との係り
受け関係を確定せず保留して構文解析処理を先に進め、
一方上記係り受け判定部が上記埋込み文との係り受け関
係を判定した名詞までで上記埋込み文が係る可能性のあ
る名詞の存在する範囲が終了したと上記名詞句範囲判定
部が判定した場合に、上記埋込み文と係り先決定部が決
定した名詞との係り受け関係を確定し解析木を作成する
ことを特徴とする。According to a second aspect of the present invention, in the first aspect of the present invention, the parse tree creating unit includes a noun for which the dependency sentence may be related to a noun for which the dependency determination unit has determined a dependency relationship with the embedded sentence. When the noun phrase range determination unit determines that the range in which the noun does not end does not determine the dependency relationship between the embedded sentence and the noun, and suspends the parsing process,
On the other hand, when the noun phrase range determination unit determines that the range in which the noun to which the embedded sentence may be concerned has ended up to the noun for which the dependency determination unit has determined the dependency relationship with the embedded sentence. And determining a dependency relationship between the embedded sentence and the noun determined by the dependency determination unit and creating a parse tree.
【０００９】請求項３の発明は、請求項２の発明におい
て解析木作成部が、係り受け判定部が埋込み文との係り
受け関係を判定した名詞までで上記埋込み文が係る可能
性のある名詞の範囲が終了しないと名詞句範囲判定部が
判定した場合であって、かつ上記係り受け関係を判定し
た名詞に上記埋込み文が係るならば該名詞に上記埋込み
文が係る可能性がある旨の情報及び上記埋込み文と名詞
との係りの属性に関する情報を解析木に付加し、上記係
り受け関係を判定した名詞に上記埋込み文が係らないな
らば上記解析木を何等操作せずに構文解析処理を先に進
めることを特徴とする。According to a third aspect of the present invention, in the second aspect of the present invention, the parse tree creation unit includes a noun whose dependency determination unit determines a dependency relationship with the embedded sentence to which the embedded sentence may relate. Is the case where the noun phrase range determination unit determines that the range is not over, and if the embedded sentence pertains to the noun for which the dependency relation has been determined, the possibility that the embedded sentence pertains to the noun Information and information on the attribute of the relationship between the embedded sentence and the noun are added to the parse tree, and if the embedded sentence does not relate to the noun for which the dependency relationship has been determined, the parsing process is performed without performing any operation on the parse tree. Is characterized by being advanced.
【００１０】請求項４の発明は、請求項１の発明におい
て名詞句範囲判定部が、係り受け判定部が埋込み文との
係り受け関係を判定した名詞の次の形態素に着目し、こ
の位置に名詞句の継続性が認められない品詞がある場合
に該名詞までで埋込み文が係る可能性のある名詞の存在
する範囲が終了したと判定し、上記名詞の次に名詞句の
継続性が認められる品詞がある場合に該名詞までで埋込
み文が係る可能性のある名詞の存在する範囲が終了しな
いと判定することを特徴とする。According to a fourth aspect of the present invention, in the first aspect of the present invention, the noun phrase range determination unit focuses on a morpheme next to the noun for which the dependency determination unit has determined a dependency relationship with an embedded sentence. If there is a part of speech where the continuity of the noun phrase is not recognized, it is determined that the range of the noun to which the embedded sentence may be related ends up to the noun, and the continuity of the noun phrase is recognized next to the above noun. If there is a part-of-speech to be performed, it is determined that the range in which the noun to which the embedded sentence may be related does not end up to the noun.
【００１１】請求項５の発明は、請求項４の発明におい
て名詞句範囲判定部が、係り受け判定部が埋込み文との
係り受け関係を判定した名詞の次の形態素に着目し、こ
の位置に格助詞がある場合に該名詞までで埋込み文が係
る可能性のある名詞の存在する範囲が終了したと判定
し、上記名詞の次に埋込み文を除く名詞句を構成する品
詞（名詞、連体詞、連体助詞、形容詞、形容動詞等）が
ある場合に該名詞までで埋込み文が係る可能性のある名
詞の存在する範囲が終了しないと判定することを特徴と
する。According to a fifth aspect of the present invention, in the fourth aspect of the present invention, the noun phrase range determination unit focuses on a morpheme next to the noun for which the dependency determination unit has determined a dependency relationship with the embedded sentence, and the position is determined at this position. When there is a case particle, it is determined that the range in which the noun to which the embedded sentence may be related to the noun is completed, and the part of speech (noun, adnominal, When there is a noun particle, an adjective, an adjective verb, etc.), it is determined that the range in which the noun to which the embedded sentence may be related does not end up to the noun.
【００１２】[0012]
【実施例】以下、本発明の実施例について図面を参照し
て説明する。図１は本発明の日本語解析装置の一実施例
を示すブロック図である。図示のように、本実施例の日
本語解析装置は、入力した日本語文に対して形態素解析
処理を行う形態素解析部１と、形態素解析部１によって
形態素解析のなされた日本語文の形態素の係り受け関係
を判断し、形態素の係り受け関係を表わす解析木を作成
する構文解析部２とを備えると共に、デ−タベ−ス７と
して日本語辞書部３、構文解析ルール集４、名詞句範囲
判定ルール集５、係り先決定ルール集６を備えてなる。Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing one embodiment of the Japanese language analysis device of the present invention. As shown in the drawing, the Japanese language analyzing apparatus according to the present embodiment includes a morphological analysis unit 1 that performs a morphological analysis process on an input Japanese sentence, and a dependency of a morpheme of the Japanese sentence subjected to morphological analysis by the morphological analysis unit 1. A syntactic analysis unit 2 for judging the relationship and creating a parse tree representing the dependency relationship of the morpheme; a Japanese dictionary unit 3 as a database 7; a syntactic analysis rule set 4; Collection 5 and a destination determination rule collection 6.
【００１３】上記構文解析部２は、形態素解析のなされ
た日本語文に対して形態素ごとにその係り受け関係を判
定する係り受け判定部２１と、係り受け判定部２１の判
断に基づいて形態素間の係り受け関係を表わす解析木を
作成する解析木作成部２２とを備え、さらに、係り受け
判定部２１が係り受け関係を判定した形態素のうち着目
中の形態素について埋込み文が係る可能性のある名詞の
存在する範囲（以下、この範囲を「名詞句の範囲」とい
う。）に関する判定を行うか否かを選択する解析木判断
部２３と、解析木判断部２３の選択に応じて名詞句の範
囲が終了したか否かを判定する名詞句範囲判定部２４
と、名詞句範囲判定部２４が名詞句の範囲が終了したと
判断した場合に埋込み文が名詞句の範囲内のいずれの名
詞に係るか決定する係り先決定部２５とを備える。The parsing unit 2 determines a dependency relationship for each morpheme in a Japanese sentence that has undergone morphological analysis, and determines a dependency relationship between morphemes based on the determination of the dependency determination unit 21. A parse tree creating unit 22 for creating an analytic tree representing a dependency relationship, and a noun whose embedded sentence may be associated with the morpheme of interest among the morphemes for which the dependency determination unit 21 has determined the dependency relationship. (Hereinafter referred to as a “range of noun phrases”). A parse tree judgment unit 23 that selects whether or not to make a judgment on a range in which the noun phrase exists. Noun phrase range determination unit 24 that determines whether or not the search has ended
And a destination determination unit 25 that determines, when the noun phrase range determination unit 24 determines that the range of the noun phrase has ended, determines which of the nouns in the range of the noun phrase the embedded sentence relates to.
【００１４】上記解析木作成部２２は、名詞句範囲判定
部２４の判定に応じて、名詞に上記埋込み文が係る可能
性がある旨の情報及び上記埋込み文と名詞との係りの属
性に関する情報を上記解析木に付加し、または上記係り
先決定部の判断及び上記解析木に付加した上記係りの属
性に関する情報に基づき解析木を作成する。なお、解析
木判断部２３は、必ずしも必須の構成要件ではなく、名
詞句範囲判定部２４において選択的に判定を行ってもよ
い。According to the determination by the noun phrase range determining unit 24, the parse tree creating unit 22 provides information indicating that the embedded sentence may be associated with the noun and information on the attribute of the relationship between the embedded sentence and the noun. Is added to the analytic tree, or an analytic tree is created based on the determination by the stakeholder determination unit and information on the attribute of the stake added to the analytic tree. Note that the parse tree determination unit 23 is not necessarily an essential component, and the noun phrase range determination unit 24 may selectively perform the determination.
【００１５】次に本実施例の作用について説明する。本
実施例に日本語文が入力されると、形態素解析部１が該
日本語文を形態素に分解し、日本語辞書部３を参照して
各形態素にその構文情報及び意味情報を付加する。日本
語辞書部３には、形態素に関し品詞の種類等の構文情報
や他の形態素との係りの可能性を制約する意味情報が形
態素ごとに分類されて格納されている。Next, the operation of this embodiment will be described. When a Japanese sentence is input to the present embodiment, the morphological analysis unit 1 decomposes the Japanese sentence into morphemes and adds syntax information and semantic information to each morpheme with reference to the Japanese dictionary unit 3. The Japanese dictionary unit 3 stores syntactic information such as the type of part of speech for morphemes and semantic information that restricts the possibility of being related to other morphemes for each morpheme.
【００１６】形態素解析部１から形態素解析のなされた
日本語文が構文解析部２に送られると、形態素が二つず
つ順番に取出され、形態素間の係り受け関係が判断され
て解析木が作成されるという処理が繰返される。When a morphologically analyzed Japanese sentence is sent from the morphological analysis unit 1 to the syntactic analysis unit 2, morphemes are sequentially extracted two by two, and the dependency relationship between the morphemes is determined to generate a parse tree. Is repeated.
【００１７】係り受け判定部２１は、形態素に付加され
た構文情報及び意味情報に基づき、構文解析ルール集４
の構文解析ルールに従って形態素間の係り受け関係を判
断する。ここで、構文解析ルール集４には、係り受け関
係の可否に関する判定法や係り方等のルールが格納され
ている。The dependency determination unit 21 generates a syntax analysis rule collection 4 based on syntax information and semantic information added to the morpheme.
Judge the dependency relationship between morphemes according to the parsing rules of. Here, the syntax analysis rule collection 4 stores rules such as a determination method regarding whether or not a dependency relationship is involved and a manner of dependency.
【００１８】解析木判断部２３は、係り受け判定部２１
によって係り受け関係を判断された形態素及びそれまで
の解析によりその形態素にかけられている部分解析木に
着目し、係りを待つ形態素が埋込み文の述語であり、着
目中の形態素が名詞である場合に、その解析結果を名詞
句範囲判定部２４へ送る。上記以外の場合は、解析結果
を直接解析木作成部２２へ送る。The parse tree determination unit 23 includes a dependency determination unit 21
Focuses on the morpheme for which the dependency relationship was determined by the morpheme and the partial parse tree applied to the morpheme by the analysis up to that point, and the morpheme waiting for the dependency is the predicate of the embedded sentence, and The analysis result is sent to the noun phrase range determination unit 24. In cases other than the above, the analysis result is directly sent to the analysis tree creating unit 22.
【００１９】名詞句範囲判定部２４は、上記着目中の形
態素である名詞を含む名詞句の範囲がその名詞までで終
了したか否かを、その名詞の次の形態素にも着目し名詞
句範囲判定ルール集５の判定ルールに従って判定する。
ここで、名詞句範囲判定ルール集５には、当該形態素に
よって名詞句の範囲が終了したか否かを判定するための
ルール（例えば、「形態素の次に格助詞があれば名詞句
の範囲が終了する」、「形態素の次に連体助詞があれば
名詞句の範囲が終了しない」等）が格納されている。The noun phrase range determination unit 24 determines whether or not the range of the noun phrase including the noun that is the morpheme of interest ends with the noun, by focusing on the next morpheme of the noun. The determination is made according to the determination rules of the determination rule collection 5.
Here, the noun phrase range determination rule collection 5 includes rules for determining whether the range of the noun phrase is terminated by the morpheme (for example, “If the case particle is followed by the morpheme, the range of the noun phrase is End, "if there is an adjunct particle following the morpheme, the range of the noun phrase will not end", etc.) are stored.
【００２０】そして、名詞句の範囲が終了したと判定し
た場合、解析結果を係り先決定部２５へ送る。名詞句の
範囲が終了しないと判定した場合、上記係りを待つ形態
素と着目中の形態素との間の係りを保留として解析結果
を解析木作成部２２へ送る。When it is determined that the range of the noun phrase has been completed, the analysis result is sent to the destination determining unit 25. If it is determined that the range of the noun phrase does not end, the relationship between the morpheme waiting for the relationship and the morpheme of interest is suspended and the analysis result is sent to the parse tree creating unit 22.
【００２１】係り先決定部２５は、上記係りを待つ形態
素である埋込み文の述語が名詞句の範囲内のいずれの名
詞に係るかを、係り先決定ルール集６の係り先決定ルー
ルに従って決定し、解析結果を解析木作成部２２へ送
る。ここで、係り先決定ルール集６には、埋込み文の種
類や名詞句の範囲内の名詞の種類、名詞間の関係等に基
づいて埋込み文の述語が名詞句の範囲内のいずれの名詞
に係るかを規定したルール（例えば、「対格と時間格の
選択においては、時間属性で他の名詞に係る時間の意味
素性をもつ名詞よりも対格で係る名詞を優先する」等）
が格納されている。The destination determining unit 25 determines which predicate of the embedded sentence, which is a morpheme waiting for the above-mentioned dependency, relates to any of the nouns in the range of the noun phrase in accordance with the destination determining rules of the destination determining rule set 6. Then, the analysis result is sent to the analysis tree creating unit 22. Here, the dependency determination rule set 6 indicates that the predicate of the embedded sentence is set to any of the nouns within the noun phrase based on the type of the embedded sentence, the type of the noun within the range of the noun phrase, the relation between the nouns, and the like. Rules that stipulate whether or not (for example, in selecting an accusative case and a temporal case, the accusative case noun takes precedence over a noun that has the temporal semantic feature of another noun in the time attribute)
Is stored.
【００２２】解析木作成部２２は、解析木判断部２３か
ら解析結果を送られた場合、係り受け判定部２１の判定
に従って形態素をかけ、解析木を作成する。また、名詞
句範囲判定部２４から解析結果を送られた場合、着目中
の名詞に係りを待つ形態素（埋込み文の述語）が係るな
らば、係り受け判定部２１の判定に基づいて、係りを待
つ形態素に係り先及びその係り先への係り方を付加情報
として付加し、構文解析処理を先に進める。一方、上記
係り受け関係を判定した名詞に係りを待つ形態素が係ら
ないならば、上記解析木を何等操作することなく構文解
析処理を先に進める。したがって、係りを待つ形態素と
着目中の形態素との間の係り受け関係の確定（解析木の
作成）は先送りにされる。When the analysis result is sent from the analysis tree determination unit 23, the analysis tree creation unit 22 applies a morpheme according to the determination of the dependency determination unit 21 and creates an analysis tree. Further, when the analysis result is sent from the noun phrase range determination unit 24, if a morpheme (a predicate of an embedded sentence) waiting for the noun concerned with the noun of interest is concerned, the dependency is determined based on the determination of the dependency determination unit 21. The morpheme to be waited for is added as a piece of additional information, and the method of connection to the morpheme, and the syntax analysis process proceeds. On the other hand, if the morpheme waiting for the change is not related to the noun for which the dependency relation has been determined, the syntax analysis process proceeds without performing any operation on the parse tree. Therefore, the determination of the dependency relationship between the morpheme waiting for the dependency and the morpheme of interest (creation of the analytic tree) is postponed.
【００２３】係り先決定部２５から解析結果を送られた
場合、着目中の形態素を係り先決定部２５によって決定
された係り先に上記の付加情報に基づいてかけ、解析木
を作成する。そして、それぞれの処理の後解析木を係り
受け判定部２１に送る。また、入力した日本語文の全て
の形態素について係り受け関係の解析が終了したとき
は、解析木を解析結果として出力する。When the analysis result is sent from the destination determination unit 25, the morpheme of interest is multiplied to the destination determined by the destination determination unit 25 based on the additional information, and an analysis tree is created. Then, after each process, the analysis tree is sent to the dependency determination unit 21. When the analysis of the dependency relation is completed for all the morphemes of the input Japanese sentence, the parse tree is output as the analysis result.
【００２４】次に、本実施例による日本語文の解析例に
ついて説明する。図２は本実施例に「本紙で経営者２０
人が選んだ今年の有望銘柄が買われている。」という文
を入力した場合の解析例を示す。本例文には、図２
（Ａ）に示すように埋込み文ａ「本紙で経営者２０人が
選んだ」と名詞句ｂ「今年の有望銘柄」とが存在し、述
語「選んだ」が名詞「今年」と名詞「有望銘柄」のいず
れに係るかが問題となる。下記の表に述語「選んだ」の
係り先を決定する手順を示す。Next, an example of analyzing a Japanese sentence according to this embodiment will be described. FIG. 2 shows that "this is the management 20
Promising brands of the year selected by people are being bought. Here is an example of the analysis when the sentence "is input. In this example sentence,
As shown in (A), there is an embedded sentence a “selected by 20 managers in the book” and a noun phrase b “promising brand of the year”, and the predicate “selected” has the noun “this year” and the noun “promising” Issue). The following table shows the procedure for determining the destination of the predicate "chosen".
【００２５】[0025]
【表１】 [Table 1]
【００２６】まず 表中の処理に示すように、述語
「選んだ」と名詞「今年」との係り受け関係が係り受け
判定部２１によって判定される。ここでは時間格として
係ると判定される。そして、名詞「今年」の後の形態素
が連体助詞「の」であることから、名詞句範囲判定部２
４によって名詞句の範囲が終了しないと判定され係りを
保留とされて、解析木作成部２２によって述語「選ん
だ」に係り先：今年， 係り方：時間格という情報が
付加される。First, as shown in the processing in the table, the dependency relationship between the predicate “chosen” and the noun “this year” is determined by the dependency determination unit 21. Here, it is determined that it is related to the temporal status. Since the morpheme after the noun “this year” is the adjunct particle “no”, the noun phrase range determination unit 2
4, it is determined that the range of the noun phrase does not end, and the relevance is suspended, and the parse tree creating unit 22 adds information of the predicate “chosen” to the predecessor: this year, the predecessor manner: temporal.
【００２７】次に、処理乃至に示すように、名詞
「今年」と連体助詞「の」、述語「選んだ」と連体助詞
「の」、連体助詞「の」と名詞「有望銘柄」のそれぞれ
の係り受け関係が係り受け判定部２１によって判定さ
れ、解析木作成部２２によって解析木が作成される。Next, as shown in the processing, the noun “this year” and the adjunct particle “no”, the predicate “chosen” and the adjunct particle “no”, and the adjunct particle “no” and the noun “promising brand” The dependency relationship is determined by the dependency determination unit 21, and the parse tree is created by the parse tree creation unit 22.
【００２８】次に、処理に示すように、述語「選ん
だ」と名詞「有望銘柄」との係り受け関係が係り受け関
係判定部２１によって判定される。ここでは対格として
係ると判定される。そして、名詞「有望銘柄」の後の形
態素が格助詞「が」であることから名詞句範囲判定部２
４によって名詞句の範囲が終了したと判定され、係り先
決定部２５によって述語「選んだ」が名詞「今年」と名
詞「有望銘柄」のいずれに係るか決定される。ここで
は、「対格と時間格の選択においては、時間属性で他の
名詞に係る時間の意味素性をもつ名詞よりも対格で係る
名詞を優先する」という係り先決定ルールにより、述語
「選んだ」は名詞「今年」には係らず名詞「有望銘柄」
に係ると決定される。そして、解析木作成部２２によっ
て述語「選んだ」は名詞「有望銘柄」に対格としてかけ
られる。この後、残りの形態素について係り受け関係が
判定され、解析木が作成される。そして、図２（Ｂ）に
示すような解析木が解析結果として出力されることとな
る。Next, as shown in the processing, the dependency relationship determining unit 21 determines the dependency relationship between the predicate "chosen" and the noun "promising brand". Here, it is determined to be accused. Then, since the morpheme after the noun “promising brand” is the case particle “ga”, the noun phrase range determination unit 2
4, it is determined that the range of the noun phrase has ended, and the destination determination unit 25 determines whether the predicate “selected” relates to the noun “this year” or the noun “promising brand”. Here, the predicate "selected" is given by the dependency determination rule that "in the selection of the accusative case and the temporal case, the noun having the temporal attribute and the semantic feature of time related to other nouns is given priority over the noun having the accusative case". Is noun "promising brand" regardless of the noun "this year"
Is determined. Then, the predicate “selected” is applied to the noun “promising brand” by the parse tree creating unit 22 as an accusative case. Thereafter, a dependency relationship is determined for the remaining morphemes, and a parse tree is created. Then, an analysis tree as shown in FIG. 2B is output as an analysis result.
【００２９】[0029]
【発明の効果】以上説明したように、本発明は、構文解
析部が名詞句範囲判定部と係り先決定部と、必要に応じ
て係り受け関係に関する情報を暫定的に解析木に付加情
報として付加する解析木作成部とを備えるため、埋込み
文が複数の名詞を含む名詞句内のいずれかの名詞に係り
かつ複数の名詞に係り得るといった構造を有する日本語
文に対しても効率よく上記埋込み文の正しい係り先を判
断することができるという効果がある。また、名詞句の
範囲が終了するまで埋込み文と名詞との係り受け関係を
確定しないため、埋込み文の正当な係り先と判断される
名詞が複数ある場合にも、それぞれの名詞との係り受け
関係に関する情報を集めた上で最も正しい係り先を判断
することができ、解析精度の向上を図ることができると
いう効果がある。As described above, according to the present invention, according to the present invention, the syntactic analysis section temporarily stores information relating to the noun phrase range determination section, the destination determination section, and, if necessary, the dependency relation as additional information in the parse tree. Since the analytic tree creation unit to be added is provided, the embedding can be efficiently performed even for a Japanese sentence having a structure in which the embedded sentence is related to any one of the noun phrases including a plurality of nouns and may be related to a plurality of nouns. This has the effect that the correct destination of the sentence can be determined. In addition, since the dependency relationship between the embedded sentence and the noun is not determined until the end of the noun phrase range, even when there are multiple nouns that are determined to be valid dependencies of the embedded sentence, the dependency with each noun is It is possible to determine the most correct participant after collecting information on the relationship, and it is possible to improve the analysis accuracy.
【図１】本発明の日本語解析装置の一実施例を示すブロ
ック図である。FIG. 1 is a block diagram showing an embodiment of a Japanese language analysis device according to the present invention.
【図２】本実施例による日本語文の解析例を示す図であ
る。FIG. 2 is a diagram illustrating an example of analyzing a Japanese sentence according to the embodiment.
１ 形態素解析部 ２ 構文解析部 ３ 日本語辞書部 ４ 構文解析ルール集 ５ 名詞句範囲判定ルール集 ６ 係り先決定ルール集 ７ デ−タベ−ス ２１ 係り受け判定部 ２２ 解析木作成部 ２３ 解析木判断部 ２４ 名詞句範囲判定部 ２５ 係り先決定部 DESCRIPTION OF SYMBOLS 1 Morphological analysis part 2 Parse analysis part 3 Japanese dictionary part 4 Parse analysis rule collection 5 Noun phrase range determination rule collection 6 Dependency determination rule collection 7 Data base 21 Dependency determination part 22 Parse tree creation part 23 Parse tree Judgment unit 24 Noun phrase range judgment unit 25 Dependency determination unit
───────────────────────────────────────────────────── フロントページの続き (58)調査した分野(Int.Cl.7，ＤＢ名) G06F 17/27 - 17/28 ──────────────────────────────────────────────────続 き Continued on the front page (58) Field surveyed (Int.Cl. 7 , DB name) G06F 17/27-17/28
Claims (5)
行なう形態素解析部と、 上記形態素解析処理のなされた日本語文の形態素の係り
受け関係を判断し、形態素の係り受け関係を表わす解析
木を作成する構文解析部と、 上記構文解析部に接続され、構文解析部が参照するデー
タベースとして、 係り受け関係の可否に関する判定法や係り方のルールが
格納された構文解析ルール集格納部、 形態素によって名詞句の範囲が終了したか否かを判定す
るルールが格納された名詞句範囲判定ルール集格納部、 及び埋込み文の種類や名詞句の範囲内の名詞の種類、名
詞間の関係等に基づいて埋込み文の述語が名詞句の範囲
内のいずれの名詞に係るかを規定したルールが格納され
た係り先決定ルール集格納部を備えてなり、入力された
日本語文の係り受け構造を機械的に解析して解析結果を
出力する日本語解析装置において、 上記構文解析部は、形態素間に係り受け関係が成立する
か否かを上記構文解析ルール集格納部を参照して判定す
る係り受け判定部と、 該係り受け判定部に接続され、係り受け判定部の判定結
果に基づいて形態素の係り受け関係を表わす解析木を作
成する解析木作成部とを備えると共に、 上記係り受け判定部に接続され、係り受け判定部で係り
受け関係を判断された形態素、及びそれまでの解析によ
りその形態素にかけられている部分解析木に着目し、着
目中の形態素への係りを待つ形態素が埋込み文の述語で
ない場合、及び着目中の形態素が名詞でない場合はその
解析結果を上記解析木作成部へ送る解析木判断部と、 上記解析木判断部に接続され、係り受け判定部で上記解
析木作成部の作成した解析木によって表わされる埋込み
文と該埋込み文の後方の名詞との係り受け関係を判定し
た際に、該名詞までで上記埋込み文が係る可能性のある
名詞の存在する範囲が終了したか否かを名詞句範囲判定
ルール集格納部を参照して判定する名詞句範囲判定部
と、 該名詞句範囲判定部に接続され、名詞句範囲判定部で上
記埋込み文が係る可能性のある名詞の範囲が終了したと
判定した場合に上記埋込み文が上記名詞の範囲内のいず
れの名詞に係るかを係り先決定ルール集格納部を参照し
て決定する係り先決定部とを備えてなり、 上記解析木作成部は、上記名詞句範囲判定部及び上記係
り先決定部に接続され、上記名詞句範囲判定部の判定に
応じて、名詞に上記埋込み文が係る可能性がある旨の情
報、及び上記埋込み文と名詞との係りの属性に関する情
報を上記解析木に付加し、または上記係り先決定部での
判定及び上記解析木に付加した上記係りの属性に関する
情報に基づき解析木を作成することを特徴とする日本語
解析装置。1. A morphological analysis unit for performing morphological analysis processing on an input Japanese sentence, and determining a morphological dependency relation of the Japanese sentence subjected to the morphological analysis processing, and forming a parse tree representing the morphological dependency relation. A parsing unit to be created; a parsing rule collection storage unit that is connected to the parsing unit and that stores a determination method regarding whether or not a dependency relationship is involved and a rule as to a relation as a database referred to by the parsing unit; A noun phrase range determination rule collection storing rules for determining whether or not the noun phrase range has ended, based on the type of embedded sentence, the type of noun within the range of the noun phrase, the relationship between nouns, etc. And a rule determination storage unit for storing rules specifying which nouns within the range of the noun phrase the predicate of the embedded sentence relates to. In Japanese analyzing apparatus for outputting an analysis result by mechanically analyze only structure, the syntax analysis unit determines whether received relates between morphemes relationship is established with reference to the parsing rule collection storage unit A dependency determining unit to be determined, and a parse tree creating unit connected to the dependency determining unit for creating a parse tree representing a morphological dependency relationship based on the determination result of the dependency determining unit. A morpheme that is connected to the reception determining unit and focuses on the morpheme for which the dependency relationship has been determined by the dependency determination unit and the partial analytic tree that has been applied to the morpheme by analysis up to that point, and waits for the morpheme being focused on If is not a predicate of an embedded sentence, and if the morpheme of interest is not a noun, a parse tree determination unit that sends the analysis result to the parse tree creation unit; and a dependency determination unit connected to the parse tree determination unit When determining the dependency relationship between the embedded sentence represented by the parse tree created by the parse tree creation unit and the noun behind the embedded sentence, the presence of a noun to which the embedded sentence may relate up to the noun A noun / phrase range determining unit that determines whether or not the range to be completed has ended by referring to the noun / phrase range determining rule collection storage unit; and the noun / phrase range determining unit, wherein the embedded sentence is connected to the noun / phrase range determining unit. Dependency determining unit that determines which of the nouns in the range of the noun the embedded sentence pertains to by referring to a destination determination rule collection storage unit when it is determined that the range of the noun that may be related ends. The analytic tree creation unit is connected to the noun phrase range determination unit and the destination determination unit, and the embedded sentence may be associated with the noun according to the determination of the noun phrase range determination unit. Information that there is, and embedded above And information about the attribute of the relationship between the noun and the noun is added to the parse tree, or a parse tree is created based on the information on the attribute of the relationship added to the parse tree determined by the deciding destination determining unit. Japanese analyzer for.
文との係り受け関係を判定した名詞までで上記埋込み文
が係る可能性のある名詞の存在する範囲が終了しないと
名詞句範囲判定部が判定した場合に、上記埋込み文と名
詞との係り受け関係を確定せず保留して構文解析処理を
先に進め、 一方上記係り受け判定部が上記埋込み文との係り受け関
係を判定した名詞までで上記埋込み文が係る可能性のあ
る名詞の存在する範囲が終了したと上記名詞句範囲判定
部が判定した場合に、上記埋込み文と係り先決定部が決
定した名詞との係り受け関係を確定し解析木を作成する
ことを特徴とする請求項１に記載の日本語解析装置。2. A parse tree creation unit, wherein a dependency determination unit determines a dependency relationship with an embedded sentence, and a noun phrase range is required until the end of a range in which a noun to which the embedded sentence may belong is present. When the determining unit determines, the dependency relationship between the embedded sentence and the noun is not determined and is retained and the syntax analysis process proceeds, while the dependency determining unit determines the dependency relationship with the embedded sentence. If the noun phrase range determination unit determines that the range in which the noun to which the embedded sentence may possibly belong is completed up to the noun, the dependency between the embedded sentence and the noun determined by the destination determination unit The Japanese language analyzing apparatus according to claim 1, wherein the relationship is determined and a parse tree is created.
文との係り受け関係を判定した名詞までで上記埋込み文
が係る可能性のある名詞の範囲が終了しないと名詞句範
囲判定部が判定した場合であって、かつ上記係り受け関
係を判定した名詞に上記埋込み文が係るならば該名詞に
上記埋込み文が係る可能性がある旨の情報及び上記埋込
み文と名詞との係りの属性に関する情報を解析木に付加
し、上記係り受け関係を判定した名詞に上記埋込み文が
係らないならば上記解析木を何等操作せずに構文解析処
理を先に進めることを特徴とする請求項２に記載の日本
語解析装置。3. The noun phrase range determination unit, wherein the parse tree creation unit determines that the range of the noun to which the embedded sentence may relate does not end up to the noun for which the dependency determination unit has determined the dependency relationship with the embedded sentence. Is determined, and if the embedded sentence pertains to the noun for which the dependency relationship has been determined, information indicating that the embedded sentence may be associated with the noun and the relationship between the embedded sentence and the noun The information relating to attributes is added to the parse tree, and if the embedded sentence does not relate to the noun for which the dependency relation has been determined, the parsing process proceeds without performing any operation on the parse tree. Japanese analysis apparatus according to 2.
込み文との係り受け関係を判定した名詞の次の形態素に
着目し、この位置に名詞句の継続性が認められない品詞
がある場合に該名詞までで埋込み文が係る可能性のある
名詞の存在する範囲が終了したと判定し、上記名詞の次
に名詞句の継続性が認められる品詞がある場合に該名詞
までで埋込み文が係る可能性のある名詞の存在する範囲
が終了しないと判定することを特徴とする請求項１に記
載の日本語解析装置。4. The noun phrase range determination unit focuses on a morpheme next to the noun for which the dependency determination unit has determined a dependency relationship with an embedded sentence, and a part of speech in which continuity of the noun phrase is not recognized at this position. In some cases, it is determined that the range in which the noun to which the embedment sentence may relate ends up to the noun, and if there is a part of speech in which continuity of the noun phrase is recognized after the above noun, the embedment is performed up to the noun. Japanese analysis apparatus according to claim 1, the range in the presence of a noun that could statement according to and judging not to terminate.
込み文との係り受け関係を判定した名詞の次の形態素に
着目し、この位置に格助詞がある場合に該名詞までで埋
込み文が係る可能性のある名詞の存在する範囲が終了し
たと判定し、上記名詞の次に埋込み文を除く名詞句を構
成する品詞がある場合に該名詞までで埋込み文が係る可
能性のある名詞の存在する範囲が終了しないと判定する
ことを特徴とする請求項４に記載の日本語解析装置。5. A noun / phrase range determining unit pays attention to a morpheme next to a noun for which a dependency determining unit has determined a dependency relationship with an embedded sentence, and if there is a case particle at this position, embedding is performed up to the noun. It is determined that the range in which the noun to which the sentence may be concerned has ended, and if there is a part-of-speech constituting a noun phrase excluding the embedded sentence next to the above noun, the embedded sentence may be affected up to the noun. Japanese analysis apparatus according to claim 4 which range the presence of nouns and judging not to terminate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP17621591A JP3251032B2 (en) | 1991-06-20 | 1991-06-20 | Japanese analyzer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP17621591A JP3251032B2 (en) | 1991-06-20 | 1991-06-20 | Japanese analyzer |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JPH04372060A JPH04372060A (en) | 1992-12-25 |
| JP3251032B2 true JP3251032B2 (en) | 2002-01-28 |
Family
ID=16009636
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP17621591A Expired - Fee Related JP3251032B2 (en) | 1991-06-20 | 1991-06-20 | Japanese analyzer |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JP3251032B2 (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5295576B2 (en) * | 2008-01-24 | 2013-09-18 | ヤフー株式会社 | Natural language analysis apparatus, natural language analysis method, and natural language analysis program |
| JP5576707B2 (en) | 2010-05-12 | 2014-08-20 | 株式会社Ｐｆｕ | Image reading apparatus and image reading method |
-
1991
- 1991-06-20 JP JP17621591A patent/JP3251032B2/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| JPH04372060A (en) | 1992-12-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP0953192B1 (en) | Natural language parser with dictionary-based part-of-speech probabilities | |
| EP0886226B1 (en) | Linguistic search system | |
| WO2001096980A2 (en) | Method and system for text analysis | |
| JPH0242572A (en) | Preparation/maintenance method for co-occurrence relation dictionary | |
| JP3251032B2 (en) | Japanese analyzer | |
| US7921126B2 (en) | Patent summarization systems and methods | |
| JP3293619B2 (en) | Japanese parsing system | |
| JP4033093B2 (en) | Natural language processing system, natural language processing method, and computer program | |
| JPH0776971B2 (en) | Document abstract creation device | |
| JP2807236B2 (en) | Morphological analysis method | |
| JP3369584B2 (en) | Japanese parser | |
| JP2003108583A (en) | Retrieving device using language sentence, retrieving system, retrieving method, program and recording medium | |
| CN108563617B (en) | Method and device for mining Chinese sentence mixed template | |
| Cybulski et al. | The use of templates and restricted english in structuring and analysis of informal requirements specifications | |
| JP3222173B2 (en) | Japanese parsing system | |
| Sang et al. | Reduction of Dutch Sentences for Automatic Subtitling. | |
| JP3358100B2 (en) | Japanese question message analysis method and device | |
| JPH0785101A (en) | Keyword extract processing unit | |
| JP2719453B2 (en) | Machine translation equipment | |
| Vogt | Passive Feedback Collection An Attempt to Debunk The Myth of Clickthroughs | |
| JP2954592B2 (en) | Language analyzer | |
| JP2912479B2 (en) | Parsing control method | |
| JP2849400B2 (en) | Language processor | |
| JPH0320866A (en) | Text base retrieval system | |
| Safadi | Crosslingual implementation of linguistic taggers using parallel corpora |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| S531 | Written request for registration of change of domicile |
Free format text: JAPANESE INTERMEDIATE CODE: R313531 |
|
| S533 | Written request for registration of change of name |
Free format text: JAPANESE INTERMEDIATE CODE: R313533 |
|
| R350 | Written notification of registration of transfer |
Free format text: JAPANESE INTERMEDIATE CODE: R350 |
|
| R250 | Receipt of annual fees |
Free format text: JAPANESE INTERMEDIATE CODE: R250 |
|
| R250 | Receipt of annual fees |
Free format text: JAPANESE INTERMEDIATE CODE: R250 |
|
| S533 | Written request for registration of change of name |
Free format text: JAPANESE INTERMEDIATE CODE: R313533 |
|
| R350 | Written notification of registration of transfer |
Free format text: JAPANESE INTERMEDIATE CODE: R350 |
|
| R250 | Receipt of annual fees |
Free format text: JAPANESE INTERMEDIATE CODE: R250 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20071116 Year of fee payment: 6 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20081116 Year of fee payment: 7 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20081116 Year of fee payment: 7 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20091116 Year of fee payment: 8 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20091116 Year of fee payment: 8 |
|
| FPAY | Renewal fee payment (event date is renewal date of database) |
Free format text: PAYMENT UNTIL: 20101116 Year of fee payment: 9 |
|
| LAPS | Cancellation because of no payment of annual fees | ||
| S533 | Written request for registration of change of name |
Free format text: JAPANESE INTERMEDIATE CODE: R313533 |
|
| R371 | Transfer withdrawn |
Free format text: JAPANESE INTERMEDIATE CODE: R371 |