JPH01145721A - Retrieval validity deciding system for document - Google Patents

Retrieval validity deciding system for document

Info

Publication number
JPH01145721A
JPH01145721A JP62303126A JP30312687A JPH01145721A JP H01145721 A JPH01145721 A JP H01145721A JP 62303126 A JP62303126 A JP 62303126A JP 30312687 A JP30312687 A JP 30312687A JP H01145721 A JPH01145721 A JP H01145721A
Authority
JP
Japan
Prior art keywords
search
document
item
importance
retrieval
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.)
Pending
Application number
JP62303126A
Other languages
Japanese (ja)
Inventor
Akira Kagami
晃 加賀美
Koichi Honma
弘一 本間
Makoto Nomi
能見 誠
Fuminobu Furumura
文伸 古村
Shinichiro Miyaoka
宮岡 伸一郎
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hitachi Ltd
Original Assignee
Hitachi Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Hitachi Ltd filed Critical Hitachi Ltd
Priority to JP62303126A priority Critical patent/JPH01145721A/en
Publication of JPH01145721A publication Critical patent/JPH01145721A/en
Pending legal-status Critical Current

Links

Landscapes

  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

PURPOSE:To realize the labor saving of a retrieval work by attaching the information for showing a degree of importance in a document, to a retrieval item given to each document. CONSTITUTION:To a retrieval item given to each document being an object to be retrieved, information for showing a degree of importance of the retrieval item to its document is attached. For instance, an editor sets words and phrases being clearly the most important as an initial key word from a terminal 3 by referring to a title, etc., of the document. As for the information for showing a degree of importance of the key word, that which has multiplied the number of times (frequency information) by which its key word appears in an extraction object area of the determined key word by such a proportional constant as the maximum value becomes 5, and thereafter, has converted it to an integer is also available. The goodness-of-fit of a retrieval document to a retriever's retrieval intention which cannot be known by only information of existence of the retrieval item is obtained as a numerical value, therefore, by outputting preferentially a document whose goodness-of-fit is high, etc., labor saving a retrieval work can be realized.

Description

【発明の詳細な説明】 〔産業上の利用分野〕 本発明は、電子計算機を用いて検索項目の論理式により
検索を行う文献検索システムに係り、特に検索結果の出
力順序制御に好適な文献の検索妥当性判定方式に関する
[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a literature search system that uses a computer to perform a search based on logical formulas for search items, and in particular to a literature search system that is suitable for controlling the output order of search results. Concerning search validity determination method.

〔従来の技術〕[Conventional technology]

従来、検索項目を用いた文献検索システムでは、[ジク
スト(JIC8T)科学技術用語シリ−ラス(1981
年)付録〔3〕検索式の作成法」において論じられてい
るように、検索者の意図する検索テーマを検索項目のプ
ール(Boolean )論理。
Conventionally, in literature search systems using search items, [JIC8T] Science and Technology Terminology Series (1981
As discussed in Appendix [3] How to Create a Search Expression, the search theme intended by the searcher is determined by pooling (Boolean) logic of search items.

すなわち、 論理和  OR 論理積  AND 論理差  AND  NOT の組合せによって表現し、予め個々の文献に付与された
検索項目の部分集合がその組合せ論理を真とする場合に
その文献が検索されたと判断し、該当する文献全てをそ
の内容の如何に関らず一様に検索結果として回答出力し
ていた6 〔発明が解決しようとする問題点〕 しかしながら、上記従来技術は個々の文献に付与される
検索項目の存否の情報だけを利用している。検索された
ものが検索者の希望にどの程度適合するかが検索者に判
らないため、検索式に該当する文献が多数出力された場
合、例えば時系列順序で並んでいる全候補文献を最初か
ら最後まで逐−読んで検索者の意図との適合度をチエツ
クしなければならず、所望の文献だけを判別して入手す
るまでにたいへんな手間がかかるという問題があった。
That is, it is expressed by a combination of logical sum OR logical product AND logical difference AND NOT, and if a subset of search items assigned to each document in advance makes the combination logic true, it is determined that the document has been searched, All relevant documents were uniformly output as search results regardless of their content.6 [Problem to be solved by the invention] However, the above-mentioned prior art uses search items assigned to individual documents. It uses only the information about the existence or non-existence of. Since the searcher does not know how well the searched items match the searcher's wishes, if a large number of documents matching the search expression are output, for example, all candidate documents arranged in chronological order may be searched from the beginning. There is a problem in that it takes a lot of effort to identify and obtain only the desired document, as the document must be read through to the end to check the degree of compatibility with the searcher's intention.

本発明の目的は、検索者の意図への適合度に応じて文献
の出力を制御し、検索作業の省力化を実現することにあ
る 〔問題点を解決するための手段〕 上記目的は、検索の対象である個々の文献に付与する検
索項目に該文献に対する該検索項目の重要度を示す情報
を付属させる手段と、ある検索項目の入力によって検索
された文献がいかなる妥当性をもって検索されたかを示
す情報を該文献に付与された該検索項目の重要度を示す
付属情報から得る手段と、該文献の検索の妥当性を示す
情報に基づき、検索された文献の出力を制御する手段を
設けることにより達成される。
An object of the present invention is to control the output of documents according to the degree of suitability to the searcher's intention, and to realize labor saving in search work. [Means for solving the problem] A means for attaching information indicating the importance of the search item to the document to a search item assigned to each document that is the target of the search, and a means for attaching information indicating the importance of the search item to the document, and a means for indicating the validity of the document retrieved by inputting a certain search item. means for obtaining information indicating the importance of the search item assigned to the document from attached information indicating the importance of the search item, and means for controlling the output of the searched document based on information indicating the validity of the search for the document. This is achieved by

〔作用〕[Effect]

検索者が検索時、すなわち検索テーマを表現する時に利
用する検索項目は真に検索出力を期待する文献中におい
ては重要な役割を果しているはずである。一方、多数の
文献の中には同じ検索項目が文献の内容を表現する度合
いが低いものも同時に存在しており、検索項目の存否の
みを利用する検索システムではこの両者とも検索にかか
り、区別もできない。これに対し1本発明では個々の文
献に付与する検索項目にその文献における重要度を示す
情報を付属させているため、これを利用すれば検索に用
いられた検索項目を付与された文献がその検索項目によ
って検索出力される場合の妥当性を調べることができる
。これにより、検索者の意図に近い文献とそうでない文
献とを区別できるので、検索者の意図に近いと考えられ
る文献を優先的に出力するなどの制御を実現できる。
The search items that searchers use when searching, that is, when expressing a search theme, should play an important role in the literature from which search output is truly expected. On the other hand, among a large number of documents, there are also cases where the same search items have a low degree of expressing the content of the documents, and a search system that uses only the presence or absence of search items will search for both, making it difficult to distinguish between them. Can not. On the other hand, in the present invention, information indicating the importance of each document is attached to the search item assigned to each document, so if this is used, the document assigned the search item used in the search can be It is possible to check the validity of search output based on search items. This makes it possible to distinguish between documents that are close to the searcher's intent and documents that are not, so it is possible to implement control such as preferentially outputting documents that are considered to be close to the searcher's intent.

〔実施例〕〔Example〕

第1図は、本発明による文献の蓄積・検索システムの一
実施例の構成図である。
FIG. 1 is a block diagram of an embodiment of a document storage/search system according to the present invention.

このシステムは蓄積系1と検索系2より構成される。This system consists of a storage system 1 and a search system 2.

まず、蓄積系1を説明する。蓄積系1では、編集者がコ
ード入力された原文献データ5から編集用端末3と対話
しながら、検索用ファイル作成部4で検索用ファイル6
を作成する。ここで検索用ファイル6とは、原文献デー
タ5に検索項目を付与したもの、あるいはこれをもう少
し加工して抄録の形にまで要約したものなどのことであ
る。第2図に検索項目の一実施例を示す、各文献には発
行日の時系列で決定される文献番号21の他に書誌検索
項目22と内容検索項目23を付与し、検索の便を図っ
ている。書誌検索項目22は表題や著者名などの客観的
情報であり、内容検索項目23はキーワードと呼ばれる
内容の記述に関する情報である。なお、各キーワードに
はそれが文献中でどの程度重要な情報か、すなわち、そ
のキーワードによって検索しようとする検索者にとって
そのキーワードを付与された文献がどの程度膜に立つの
かを示すための重要度24なる情報が付属している0例
えば、文献1において、キーワード「電子」は重要度5
であり、必要不可欠な情報と言えるが、−カキ−ワード
r回路」は、重要度2であるため、文献の内容をキーワ
ード「電子」はどは表現していないことがわかる。とこ
ろで、これらキーワード、及びその重要度の作成を人手
に任すと ■ 手間がかかりすぎる。
First, the storage system 1 will be explained. In the storage system 1, the editor creates a search file 6 in the search file creation unit 4 while interacting with the editing terminal 3 from the original document data 5 into which the code has been input.
Create. Here, the search file 6 is the original document data 5 with search items added, or a file that has been further processed and summarized in the form of an abstract. Figure 2 shows an example of the search items.In addition to the document number 21, which is determined in chronological order of publication date, each document is given a bibliographic search item 22 and a content search item 23 to facilitate the search. ing. Bibliographic search items 22 are objective information such as titles and author names, and content search items 23 are information related to content descriptions called keywords. Furthermore, each keyword has a level of importance that indicates how important the information is in the literature, that is, how important the literature to which that keyword is assigned stands out to the searcher who is trying to search using that keyword. For example, in document 1, the keyword "electronic" has an importance level of 5.
Although it can be said to be essential information, since the importance level of ``-kaki-word r circuit'' is 2, it can be seen that the content of the document is not expressed by the keyword ``electronic''. By the way, if the creation of these keywords and their importance levels is left to humans, ■ it will take too much time and effort.

■ 個人差が生ずる。■ Individual differences occur.

等の問題が起るため1人間の介在をできるだけ少くする
アルゴリズムを採用する。それを第3図で説明する。ブ
ロック31で入力された文献5は、まずどの領域からキ
ーワードを抽出するががブロック32で編集者の端末3
がらの指定により決定される。次に編集者はブロック3
3において文献の表題等を参考にして最も重要であるこ
とが明白な語句を初期キーワードとして端末3より設定
する。ブロック34は直前に設定された(n次)キーワ
ードを含む文をブロック32で決定されたキーワードの
抽出対象領域中に探し出し、探し出された文に対しブロ
ック35が構文解析処理を施して、9次キーワードと構
文上密接な関係例えば、同格、所有、係り受は等の関係
にある未設定語を見つけ出し、ブロック36が(n+1
)次キーワードとして追加設定する。この際、その重要
度は単純にループカウンタnを利用して、5− nとす
る。ブロック34からブロック36までの処理はブロッ
ク37のループカウンタのインクリメントと、ブロック
38の判定処理で繰返しの制御を行う。
Since such problems occur, an algorithm is adopted that minimizes human intervention as much as possible. This will be explained with reference to FIG. For the document 5 input in block 31, it is determined in block 32 from which area keywords are extracted.
Determined by the designation. Next, the editor blocks 3
In step 3, a phrase that is clearly the most important is set from the terminal 3 as an initial keyword by referring to the title of the document, etc. Block 34 searches for a sentence containing the (nth) keyword set immediately before in the keyword extraction target area determined in block 32, and block 35 performs syntax analysis on the searched sentence. Block 36 finds unset words that have a close syntactic relationship with the next keyword, such as apposition, possession, dependency, etc.
)Additionally set as the next keyword. At this time, the importance level is set to 5-n simply by using the loop counter n. The processing from block 34 to block 36 is repeatedly controlled by incrementing a loop counter in block 37 and determining processing in block 38.

なお、上記キーワードの重要度を示す情報は、ブロック
32で決定されたキーワードの抽出対象領域中にそのキ
ーワードが出現する回数(頻度情報)に最大値5となる
ような比例定数をかけた後整数化したものを利用しても
よい。
Note that the information indicating the importance of the keyword is an integer obtained by multiplying the number of times the keyword appears in the keyword extraction target area determined in block 32 (frequency information) by a proportionality constant such that the maximum value is 5. You may use the converted version.

次に、検索系2を説明する。検索系2では、蓄積系1で
作成された検索用ファイル6を用いて検索を行う、検索
者は検索用端末9に検索項目の論理式で表現した検索テ
ーマを入力する0例えば、電子顕微鏡に関する文献調査
を行いたい場合は、「電子」、「顕微鏡」なる2つの検
索項目の論理積として、論理式 %式%(1) を入力する。
Next, search system 2 will be explained. In the search system 2, a search is performed using the search file 6 created in the storage system 1.The searcher inputs a search theme expressed by a logical formula of the search item into the search terminal 9. If you wish to conduct a literature search, enter the logical formula % formula % (1) as the logical product of the two search items ``electron'' and ``microscope.''

検索実行部7は「電子」と「顕微鏡」の2つの検索項目
が付与された文献を選択するが、一般にその全てが電子
顕微鏡を主題にしている文献ではないため、重要度24
を利用して検索文献出力10を制御する0例えば第2図
に示した文献1と文献100が論理式(1)を満すが、
重要度24を見ると文献1が「電子」の重要度5.「顕
微鏡」の重要度5の計10であるのに対し1文献100
は[電子」の重要度2.「顕微鏡」の重要度2の計4で
あるので1文献1の方が検索テーマすなわち検索者の検
索意図により近い文献であると判断できる。そこで、文
献1を文献100より先に出力したり1文献100の出
力をやめたりすれば。
The search execution unit 7 selects documents to which the two search items "electron" and "microscope" are assigned, but generally not all of them are documents with an electron microscope as a subject, so the importance level is 24.
For example, document 1 and document 100 shown in FIG. 2 satisfy the logical formula (1),
Looking at the importance level 24, document 1 has an "electronic" importance level of 5. ``Microscope'' has an importance level of 5 and a total of 10, while 1 document has 100.
The importance of [electronic] is 2. Since the importance level of "microscope" is 2, the total is 4, so it can be determined that 1 document 1 is a document that is closer to the search theme, that is, the search intention of the searcher. Therefore, if you output Document 1 before Document 100 or stop outputting Document 1 100.

検索者の負担を減らすことができる。This can reduce the burden on searchers.

ところで、文献によっては電子顕微鏡をエレクトロンマ
イクロスコープと表現している可能性もあるため、この
種の同義語、類似語への拡大解釈も検索実行部7がシソ
ーラス8を利用して行う。
Incidentally, since an electron microscope may be expressed as an electron microscope depending on the literature, the search execution unit 7 also uses the thesaurus 8 to expand the interpretation to synonyms and similar words of this kind.

また、検索者自身が検索時に、検索式に用いる検索項目
に重み付けすることもできる。例えば、次式 %式%(2) のように「電子」に重み4.「顕微鏡」に重み5を与え
てやれば、検索意図をより明確に表現でき、かつ、文献
の検索妥当性もより精密になる。この場合、第2図に示
した文献1の検索妥当性は、検索項目の重みと文献中の
重要度により。
Furthermore, the searcher himself or herself can weight the search items used in the search formula. For example, as in the following formula % formula % (2), "electron" is weighted 4. If a weight of 5 is given to "microscope", the search intention can be expressed more clearly, and the validity of the search for documents can be made more precise. In this case, the validity of the search for document 1 shown in FIG. 2 depends on the weight of the search item and the degree of importance in the document.

検索妥当性(「電子J AND r顕微鏡」)=「電子
J重要度5×重み4 +「顕微鏡」重要度5×重み5 =45 と計算される。以上は、検索式を検索項目の論理和を用
いて表現しなければならない時などに、より有効な手段
となり得る。もちろん、検索項目の重み付けが必要ない
ときは省略も可能であり、その場合はすべてデフォルト
値5が与えられる。
Search validity ("electron J AND r microscope") = "electron J importance 5 x weight 4 + "microscope" importance 5 x weight 5 = 45. The above method can be a more effective means when a search expression must be expressed using a logical sum of search items. Of course, if the weighting of search items is not necessary, it can be omitted, and in that case, a default value of 5 is given to all weights.

〔発明の効果〕〔Effect of the invention〕

本発明によれば、検索項目の存否の情報だけでは判らな
い、検索者の検索意図に対する検索文献の適合度が数値
として得られるので、適合度の高い文献から優先出力す
るなどにより検索作業の省力化を実現することができる
According to the present invention, the degree of suitability of searched documents to the searcher's search intention, which cannot be determined only from information on the presence or absence of search items, can be obtained as a numerical value, so the search work can be saved by preferentially outputting documents with a high degree of suitability. can be realized.

【図面の簡単な説明】[Brief explanation of the drawing]

第1図は本発明による文献の蓄積・検索システムの一実
施例の構成図、第2図は文献に付与される検索項目の一
実施例の説明図、第3図は文献中から内容の記述に関す
るキーワード及びその文献における重要度を示す情報を
取り出すアルゴリズ鳩 2 口
Fig. 1 is a configuration diagram of an embodiment of the document storage/search system according to the present invention, Fig. 2 is an explanatory diagram of an embodiment of search items added to documents, and Fig. 3 is a description of contents from documents. An algorithm for extracting information indicating keywords and their importance in literature.

Claims (1)

【特許請求の範囲】 1、少なくとも1つ以上の検索項目を入力して、文献デ
ータベースから所望の文献を取り出す文献検索システム
において、個々の文献に付与する検索項目に該文献に対
する該検索項目の重要度を示す情報を付属させる手段と
、ある検索項目の入力によつて検索された文献がいかな
る妥当性をもつて検索されたかを示す情報を該文献に付
与された該検索項目の重要度を示す付属情報から得る手
段を有することを特徴とする文献の検索妥当性判定方式
。 2、上記文献の検索妥当性を示す情報に基き、検索され
た文献の出力を制御することを特徴とする第1項記載の
文献の検索妥当性判定方式。 3、上記個々の文献に付与する検索項目に該文献に対す
る該検索項目の重要度を示す情報を付属させる手段にお
いて、まず文献の中から最も重要であると考えられる語
句を初期検索項目として少なくとも1つ設定する第1の
ステップと、該文献を構文解析することにより該初期検
索項目と関連する語句を抽出しその関連の程度から該文
献に対する該語句の重要度を示す情報を作成し付属させ
た後該語句を検索用項目として追加設定する第2のステ
ップと、必要に応じて上記第1と第2のステップを繰返
す手段とを有することを特徴とする第1項記載の文献の
検索妥当性判定方式。 4、上記初期検索項目の設定において、文献名に含まれ
る語句を用いることを特徴とする第1項記載の文献の検
索妥当性判定方式。 5、第1項記載の個々の文献に付与する検索項目に該文
献に対する該検索項目の重要度を示す情報を付属させる
手段において、該検索項目の該文書における出現頻度を
利用して該検索項目の重要度を示す情報を作成すること
を特徴とする第1項記載の文献の検索妥当性判定方式。 6、第1項記載の検索項目の入力において、検索者の判
断により該検索項目に重み付けして入力することを特徴
とする第1項記載の文献の検索妥当性判定方式。
[Claims] 1. In a literature search system in which a desired document is retrieved from a literature database by inputting at least one search item, the importance of the search item for the document is added to the search item assigned to each document. means for attaching information indicating the importance of the search item given to the document, and information indicating the validity of the document searched by inputting a certain search item. A document search validity determination method characterized by having a means for obtaining information from attached information. 2. The document search validity determination method according to item 1, wherein the output of the retrieved documents is controlled based on the information indicating the search validity of the documents. 3. In the means for attaching information indicating the importance of the search item to each document to the search item described above, first, at least one word or phrase considered to be the most important from among the documents is added as an initial search item. The first step is to analyze the literature to extract terms related to the initial search item, and create and attach information indicating the importance of the term to the document based on the degree of relationship. The search validity of documents as described in item 1, further comprising a second step of additionally setting the word/phrase as a search item, and means for repeating the first and second steps as necessary. Judgment method. 4. The document search validity determination method according to item 1, wherein words included in document names are used in setting the initial search items. 5. In the means for attaching information indicating the importance of the search item to the document to the search item described in paragraph 1, the search item is added to the search item by using the frequency of appearance of the search item in the document. 2. The document search validity determination method according to item 1, wherein information indicating the importance of the document is created. 6. The literature search validity determination method as set forth in item 1, characterized in that when inputting the search items described in item 1, the search items are weighted and input according to the searcher's judgment.
JP62303126A 1987-12-02 1987-12-02 Retrieval validity deciding system for document Pending JPH01145721A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
JP62303126A JPH01145721A (en) 1987-12-02 1987-12-02 Retrieval validity deciding system for document

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
JP62303126A JPH01145721A (en) 1987-12-02 1987-12-02 Retrieval validity deciding system for document

Publications (1)

Publication Number Publication Date
JPH01145721A true JPH01145721A (en) 1989-06-07

Family

ID=17917195

Family Applications (1)

Application Number Title Priority Date Filing Date
JP62303126A Pending JPH01145721A (en) 1987-12-02 1987-12-02 Retrieval validity deciding system for document

Country Status (1)

Country Link
JP (1) JPH01145721A (en)

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH03129472A (en) * 1989-07-31 1991-06-03 Ricoh Co Ltd Processing method for document retrieving device
JPH04262460A (en) * 1991-02-15 1992-09-17 Ricoh Co Ltd Information retrieval device
JPH05101107A (en) * 1991-10-07 1993-04-23 Hitachi Ltd Device and method for narrowed-down data retrieval using adaption rate
JPH05158991A (en) * 1991-12-02 1993-06-25 Mitsubishi Electric Corp Information retrieval system
JPH06176069A (en) * 1992-12-02 1994-06-24 Dainippon Printing Co Ltd Display device for retrieving result of character string
JPH06176065A (en) * 1992-12-02 1994-06-24 Dainippon Printing Co Ltd Retrieving device for scientific paper data
JPH07160727A (en) * 1993-12-06 1995-06-23 Fujitsu Ltd Electronic manual display method
JPH09114847A (en) * 1995-10-16 1997-05-02 Fuji Xerox Co Ltd Information processor
JPH09510811A (en) * 1995-01-11 1997-10-28 フィリップス エレクトロニクス ネムローゼ フェンノートシャップ User interface for full text document search
JPH10105571A (en) * 1996-10-02 1998-04-24 Hitachi Ltd Retrieval system

Cited By (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH03129472A (en) * 1989-07-31 1991-06-03 Ricoh Co Ltd Processing method for document retrieving device
JPH04262460A (en) * 1991-02-15 1992-09-17 Ricoh Co Ltd Information retrieval device
JPH05101107A (en) * 1991-10-07 1993-04-23 Hitachi Ltd Device and method for narrowed-down data retrieval using adaption rate
JPH05158991A (en) * 1991-12-02 1993-06-25 Mitsubishi Electric Corp Information retrieval system
JPH06176069A (en) * 1992-12-02 1994-06-24 Dainippon Printing Co Ltd Display device for retrieving result of character string
JPH06176065A (en) * 1992-12-02 1994-06-24 Dainippon Printing Co Ltd Retrieving device for scientific paper data
JPH07160727A (en) * 1993-12-06 1995-06-23 Fujitsu Ltd Electronic manual display method
JPH09510811A (en) * 1995-01-11 1997-10-28 フィリップス エレクトロニクス ネムローゼ フェンノートシャップ User interface for full text document search
JP2004005742A (en) * 1995-01-11 2004-01-08 Koninkl Philips Electronics Nv User interface for document full-text search
JPH09114847A (en) * 1995-10-16 1997-05-02 Fuji Xerox Co Ltd Information processor
JPH10105571A (en) * 1996-10-02 1998-04-24 Hitachi Ltd Retrieval system

Similar Documents

Publication Publication Date Title
JP3099756B2 (en) Document processing device, word extraction device, and word extraction method
US7440947B2 (en) System and method for identifying query-relevant keywords in documents with latent semantic analysis
JP4944405B2 (en) Phrase-based indexing method in information retrieval system
US6523030B1 (en) Sort system for merging database entries
US20060190446A1 (en) Web search system and method thereof
JPH0424869A (en) Document processing system
EP1604309A2 (en) Corpus clustering, confidence refinement, and ranking for geographic text search and information retrieval
JP2006048683A (en) Phrase identification method in information retrieval system
CN112000783B (en) Patent recommendation method, device and equipment based on text similarity analysis and storage medium
US6278990B1 (en) Sort system for text retrieval
JP2669601B2 (en) Information retrieval method and system
JP2003281186A (en) Example base retrieval method and retrieval system for determining similarity
JP4857448B2 (en) Information retrieval apparatus and program using multiple meanings
JP3584848B2 (en) Document processing device, item search device, and item search method
JPH10260972A (en) Relative document retrieval device and record medium where relative document retrieving program is recorded
JPH01145721A (en) Retrieval validity deciding system for document
JP5869948B2 (en) Passage dividing method, apparatus, and program
Singla et al. A novel approach for document ranking in digital libraries using extractive summarization
JPH0773197A (en) Supporting system for preparing different notation word dictionary
JPH03294963A (en) Document retrieving device
JP2773682B2 (en) Applicable feedback device
Gokhan et al. GUSUM: graph-based unsupervised summarization using sentence features scoring and sentence-BERT
JP2004342016A (en) Information retrieval program and medium having information retrieval program recorded thereon
JP2002117043A (en) Device and method for document retrieval, and recording medium with recorded program for implementing the same method
JP2003108582A (en) Synonym extracting method and document retrieving device