JP3068397B2 - Document management device - Google Patents
Document management deviceInfo
- Publication number
- JP3068397B2 JP3068397B2 JP5330306A JP33030693A JP3068397B2 JP 3068397 B2 JP3068397 B2 JP 3068397B2 JP 5330306 A JP5330306 A JP 5330306A JP 33030693 A JP33030693 A JP 33030693A JP 3068397 B2 JP3068397 B2 JP 3068397B2
- Authority
- JP
- Japan
- Prior art keywords
- document
- distance
- frequency
- documents
- search
- 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
- 239000013598 vector Substances 0.000 description 13
- 238000007726 management method Methods 0.000 description 7
- 150000001875 compounds Chemical class 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 238000010187 selection method Methods 0.000 description 2
- 239000000470 constituent Substances 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Description
【０００１】[0001]
【産業上の利用分野】本発明は、計算機装置における文
書管理、検索のための文書分類・文書照合を行なう文書
管理法に関する。BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document management method for performing document classification and document collation for document management and retrieval in a computer device.
【０００２】[0002]
【従来の技術】従来の文書管理装置では、文書内の単語
や文書に付与されたキーワードの存在ベクトルによっ
て、文書間の距離を算定していた。しかし、キーワード
の特定法の問題や単語の語切りの曖昧性があるため、こ
れらの頻度やベクトルは文書自身を特定するのに不十分
であった。例えば、特開昭６３−３１２１０７号公報に
記載されている文書分類装置では、キーワードの出現頻
度を用いてデータベース中でのキーワードの自己情報量
を求め、その情報量を元に各文書の特徴量を求め更に文
書間の距離を計算している。しかし、キーワードの選択
法により、あるテキストについてはそのテキストを特定
するのに充分なキーワードがない可能性もあり、重要語
彙として複合語等がある場合は複合語をキーワードとす
るかその複合語を構成する各々の語をキーワードとする
かによって、文書の位置付けが変化してしまうという問
題点がある。2. Description of the Related Art In a conventional document management apparatus, a distance between documents is calculated based on an existence vector of a word in the document or a keyword assigned to the document. However, due to the problem of how to specify keywords and the ambiguity of word segmentation, these frequencies and vectors were insufficient to identify the document itself. For example, in a document classification device described in Japanese Patent Application Laid-Open No. 63-312107, the amount of self-information of a keyword in a database is obtained using the frequency of appearance of the keyword, and the characteristic amount of each document is determined based on the amount of information. Is calculated and the distance between the documents is calculated. However, depending on the keyword selection method, there may not be enough keywords for a certain text to specify the text. If there are compound words etc. as important vocabulary, use compound words as keywords or replace the compound words. There is a problem that the position of a document changes depending on whether each of the constituent words is used as a keyword.
【０００３】[0003]
【発明が解決しようとする課題】この従来の文書管理装
置では、文書を検索／分類するための文書間の距離を、
各文書内のキーワードを元に計算しており、文書の特徴
を捉えるのに、キーワードの選択法により捉え方が変化
するという問題点があり、キーワードの選択が困難なた
め、全文書を均等に捉えられないという問題点があっ
た。また、文書が新たに入力された場合も、それまでの
文書を分類するために用意していたキーワードでその文
書が充分特徴付けられるとは限らず、もし新たにキーワ
ードを追加した場合は、そのキーワードの抽出および距
離計算を全文書に対して実行しなけらばならない等の問
題がある。In this conventional document management apparatus, the distance between documents for searching / classifying documents is determined by:
The calculation is based on the keywords in each document, and there is a problem that the way of selecting the keywords changes depending on the keyword selection method, and it is difficult to select keywords. There was a problem that it could not be caught. Also, even when a new document is entered, the keyword prepared for classifying the previous document is not always sufficiently characterized by the keyword, and if a new keyword is added, There are problems such as that keyword extraction and distance calculation must be performed for all documents.
【０００４】[0004]
【課題を解決するための手段】本発明の文書管理装置
は、文書を検索するための文書検索手段や文書を分類整
理する文書分類手段および文書を保存するための文書保
存装置を備えた文書管理装置において、 個々の文書の
全出現文字の出現頻度を計算する文字出現頻度計算手段
と、前記文字出現頻度計算手段で計算された各文書の特
徴を表す出現頻度を元に前記文書検索手段や前記文書分
類手段によって指定された文書と前記文書保存装置内の
文書との距離を計算しその距離を文書検索や文書分類に
利用させるため前記文書検索装置や前記文書分類手段へ
伝える文書間距離計算手段とを有することを特徴とす
る。SUMMARY OF THE INVENTION A document management apparatus according to the present invention comprises a document search unit for searching for a document, a document classification unit for classifying and organizing documents, and a document storage unit for storing documents.
In a document management device with a storage device,
Character appearance frequency calculation means for calculating the appearance frequency of all the appearance characters; and the document search means and the document segment based on the appearance frequency representing the characteristics of each document calculated by the character appearance frequency calculation means.
The document specified by the class means and the
In order to calculate a distance from a document and use the distance for document search and document classification, the document search apparatus and the document classification means are provided with an inter-document distance calculation means.
【０００５】[0005]
【作用】本発明では、テキスト中の全出現文字の出現頻
度ベクトルやその分布曲線をテキストのインデクスとし
てテキスト間の距離を定義することで、全文書を均等に
特徴付け、文書を他文書と弁別することを可能としてい
る。また、出現頻度ベクトルやその分布曲線上に定義さ
れる加法演算を用いることにより、インデクスの合成が
可能なテキストのインデクスとしてテキスト間の距離を
定義できる。According to the present invention, all documents are equally characterized by defining the distance between texts by defining the appearance frequency vector of all the characters appearing in the text and the distribution curve thereof as the text index, and distinguishing the documents from other documents. It is possible to do. In addition, by using an appearance frequency vector and an addition operation defined on a distribution curve thereof, a distance between texts can be defined as an index of a text that can be combined with the index.
【０００６】[0006]
【実施例】次に、本発明について図面を参照して説明す
る。図１は本発明の一実施例を示すブロック図である。
図１を参照すると、本発明の実施例は、文書を検索する
ための文書検索手段１や文書を分類整理する文書分類手
段２および文書を保存するための文書保存装置５を備え
た文書管理装置において、文書内の全出現文字の出現頻
度を計算する文字出現頻度計算手段４と、前記文字出現
頻度計算手段４で計算された各文書の特徴を表す頻度情
報を元に前記文書検索手段１や前記文書分類手段２によ
って要求された前記文書保存装置５内の全文書間の距離
を計算しその距離を文書検索や文書分類に利用させるた
め前記文書検索装置１や前記文書分類手段２へ伝える文
書間距離計算手段３とから構成される。Next, the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing one embodiment of the present invention.
Referring to FIG. 1, an embodiment of the present invention provides a document management apparatus including a document search unit 1 for searching for a document, a document classification unit 2 for classifying and organizing documents, and a document storage device 5 for storing documents. A character appearance frequency calculating means 4 for calculating the appearance frequency of all the characters appearing in the document, and the document search means 1 based on the frequency information representing the characteristics of each document calculated by the character appearance frequency calculating means 4. By the document classification means 2
Calculating the distance between all the documents in the document storage device 5 requested by the user and transmitting the distance to the document search device 1 and the document classification device 2 to use the distance for document search and document classification. And 3.
【０００７】文書検索手段１や文書分類手段２に対し、
文書の検索／分類の要求が入力された際に、これらの手
段は文書の特徴量（文字の出現頻度）を文字頻度計算手
段４に要求する。文字頻度計算手段４は、文書保存装置
５より文書を取り出し、文書中の文字の出現回数（頻
度）を縦軸にして、これを頻度ベクトルとして保持する
か、ｎ次の多項式の連続関数として保持する。この出現
頻度を元に、文書間距離計算手段３は文書の照合をこの
ベクトルの形の近さとして定義する（最小二乗距離でも
良い）。最小二乗距離としての文字の出現頻度でベクト
ルを定義した場合、その距離は以下の式となる。For the document search means 1 and the document classification means 2,
When search / classification request of the document is input, these means are required characteristics of the document (the frequency of appearance of characters) in the character frequency calculation unit 4. The character frequency calculation means 4 retrieves the document from the document storage device 5, sets the number of appearances (frequency) of the character in the document as the vertical axis, and stores this as a frequency vector or as a continuous function of an n-th order polynomial. I do. Based on the appearance frequency, the inter-document distance calculation means 3 defines the collation of the document as the closeness of the shape of the vector (least square distance may be used). When a vector is defined by the appearance frequency of a character as the least square distance, the distance is represented by the following equation.
【０００８】 [0008]
【０００９】また、文書Ａと文書Ｂの頻度ベクトルを文
書長に関してその出現頻度補正や正規化を行って足した
頻度ベクトルは文書Ａの内容と文書Ｂの内容を含む文書
Ｘの頻度ベクトルになるとかんがえ、その文書Ｘに関連
するテキストを加法合成された頻度ベクトルＮａｂを使
って照合する。Further, the frequency vector obtained by adding the frequency vectors of the document A and the document B by correcting the appearance frequency of the document length and normalizing the same becomes the frequency vector of the document X including the content of the document A and the content of the document B. Considering this, the text associated with the document X is collated using the frequency vector Nab that is additively synthesized.
【００１０】 [0010]
【００１１】これらの文書間距離を用いて、文書検索手
段１は文書検索を行ない、文書分類手段２は文書分類を
行ない、その結果を出力する。Using these inter-document distances, the document search means 1 performs a document search, and the document classification means 2 performs a document classification, and outputs the result.
【００１２】[0012]
【発明の効果】本発明では、テキスト中の出現文字の出
現頻度ベクトルやその分布曲線をテキストのインデクス
としてテキスト間の距離を定義することで、全文書を均
等に特徴付け、文書を他文書と弁別することが可能とな
っている。また、出現頻度ベクトルやその分布曲線上に
定義される加法演算を文書距離に適用することにより、
インデクスの合成が可能なテキストのインデクスとして
テキスト間の距離が定義されている。According to the present invention, all documents are uniformly characterized by defining the appearance frequency vector of the appearance character in the text and the distribution curve thereof as the index of the text to define the distance between the texts. Documents can be distinguished from other documents. Also, by applying an addition operation defined on the appearance frequency vector and its distribution curve to the document distance,
A distance between texts is defined as an index of a text that can be combined with the index.
【図１】本発明の一実施例を示すブロック図である。FIG. 1 is a block diagram showing one embodiment of the present invention.
１ 文書検索手段 ２ 文書分類手段 ３ 文書間距離計算手段 ４ 文字出現頻度計算手段 ５ 文書保存装置DESCRIPTION OF SYMBOLS 1 Document search means 2 Document classification means 3 Inter-document distance calculation means 4 Character appearance frequency calculation means 5 Document storage device
フロントページの続き (56)参考文献 特開 平２−158871（ＪＰ，Ａ) 特開 平５−54037（ＪＰ，Ａ) 特開 平３−48367（ＪＰ，Ａ) 特開 平４−239965（ＪＰ，Ａ)Continuation of front page (56) References JP-A-2-158871 (JP, A) JP-A-5-54037 (JP, A) JP-A-3-48367 (JP, A) JP-A-4-239965 (JP) , A)
Claims (1)
を分類整理する文書分類手段および文書を保存するため
の文書保存装置を備えた文書管理装置において、 個々の文書の全出現文字の出現頻度を計算する文字出現
頻度計算手段と、 前記文字出現頻度計算手段で計算された各文書の特徴を
表す出現頻度を元に前記文書検索手段や前記文書分類手
段によって指定された文書と前記文書保存装置内の文書
との距離を計算しその距離を文書検索や文書分類に利用
させるため前記文書検索装置や前記文書分類手段へ伝え
る文書間距離計算手段とを有することを特徴とする文書
管理装置。1. A document search means for retrieving a document, a document classification means for classifying and organizing documents, and for storing a document.
In the document management apparatus having a document storage device, character appearance for calculating the frequency of occurrence of all occurrences character of the individual documents
Frequency calculating means, and the document search means and the document classifying means based on the appearance frequency representing the characteristics of each document calculated by the character appearance frequency calculating means.
Document specified by column and document in said document storage device
Document management apparatus characterized by having the document distance calculation means for transmitting to the document search apparatus and the document classification unit for distance was calculated to use the distance to the document search and document classification with.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP5330306A JP3068397B2 (en) | 1993-12-27 | 1993-12-27 | Document management device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP5330306A JP3068397B2 (en) | 1993-12-27 | 1993-12-27 | Document management device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JPH07192002A JPH07192002A (en) | 1995-07-28 |
| JP3068397B2 true JP3068397B2 (en) | 2000-07-24 |
Family
ID=18231168
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP5330306A Expired - Fee Related JP3068397B2 (en) | 1993-12-27 | 1993-12-27 | Document management device |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JP3068397B2 (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3582297B2 (en) * | 1997-05-09 | 2004-10-27 | 日本電信電話株式会社 | Document classification method and apparatus, and storage medium storing document classification program |
| JP2002222083A (en) | 2001-01-29 | 2002-08-09 | Fujitsu Ltd | Case storage device and method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH0554037A (en) * | 1991-08-28 | 1993-03-05 | Fujitsu Ltd | Document classifying system |
-
1993
- 1993-12-27 JP JP5330306A patent/JP3068397B2/en not_active Expired - Fee Related
Also Published As
| Publication number | Publication date |
|---|---|
| JPH07192002A (en) | 1995-07-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US5544049A (en) | Method for performing a search of a plurality of documents for similarity to a plurality of query words | |
| US6862602B2 (en) | System and method for rapidly identifying the existence and location of an item in a file | |
| US6212525B1 (en) | Hash-based system and method with primary and secondary hash functions for rapidly identifying the existence and location of an item in a file | |
| US6826576B2 (en) | Very-large-scale automatic categorizer for web content | |
| JP3132738B2 (en) | Text search method | |
| US8626492B1 (en) | Semantic unit recognition | |
| US5812998A (en) | Similarity searching of sub-structured databases | |
| KR20010015368A (en) | A method of retrieving data and a data retrieving apparatus | |
| JP2003167914A (en) | Multimedia information search method, program, recording medium and system | |
| JP2669601B2 (en) | Information retrieval method and system | |
| JPH11102377A (en) | Method and device for retrieving document from data base | |
| JP3829506B2 (en) | Document classification method, document classification apparatus, and recording medium on which document classification processing program is recorded | |
| JP2000222418A (en) | Database search method and apparatus | |
| JP3068397B2 (en) | Document management device | |
| JPH08314966A (en) | Document creating apparatus index creation method and document searching apparatus | |
| JP3081093B2 (en) | Index creation method and apparatus and document search apparatus | |
| JP4426893B2 (en) | Document search method, document search program, and document search apparatus for executing the same | |
| JP3772401B2 (en) | Document classification device | |
| JP3288063B2 (en) | Variable length data storage and reference system | |
| JP3902825B2 (en) | Document search system and method | |
| JP2993539B2 (en) | Database search system and method | |
| JPH10232871A (en) | Search device | |
| JP3019286B2 (en) | Document search device | |
| EP0592402B1 (en) | A text management system | |
| JPH09160908A (en) | Document processing apparatus and method, storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| LAPS | Cancellation because of no payment of annual fees |