Abstract: Full text data is divided into words to generate word separation data. All character strings are extracted from the full text data, each character string including N characters. The word separation and position data is attached to each character string to generate index data. In word retrieving, character and segmentation agreement between query data and all character strings is checked. Word retrieving and/or character string retrieving are effected according to a selection command. The word separation data may include leading or trailing end data. In the word retrieving mode, the leading end of the first character and the trailing end of the last character are checked but the intermediate portion is not checked. Continuity of retrieve character strings is checked with reference to position data thereof. The word retrieving mode includes a number of modes including the completion agreement mode.

Abstract: A character recognizing apparatus has a post-processing unit which makes character strings including a plurality of conversion candidates, respectively, made by a character recognizing unit, and a full text searching unit performs a full text search for the character strings in a plurality of documents having been converted into text data, whereby the post-processing unit determines a correct character on the basis of results of the search to correct misrecognition.

Abstract: All two-character chains including two general characters and all three-character chains including one special character between two general characters are detected from a registration character string in which a large number of special characters not having any meaning are frequently arranged, or all two-character chains including two general or symbolic characters are detected from a converted registration character string produced by changing each special character of the registration character string to one type of symbolic character determined in correspondence to one general character adjacent to the special character. Also, occurrence frequencies of the general or symbolic characters of each chain are counted and stored in a recording medium with the registration character chains.

Abstract: A data searching apparatus is disclosed which comprises: a request receiving portion for receiving a searching request including searching data and a registering request; a record number providing portion responsive to the registering request for providing record number data to each record which is a unit of data to be registered; a relatively-numbering portion for grouping the record number data into groups and converting the record number data into group number data and relative record number data within each of the groups; an index file generation portion for generating an index file from the data and the relative record number data; an index file storing portion for storing the index file; an index file searching portion for searching one of records indicated by the searching data in accordance with the index file in the index file storing portion in response to the searching request from the request receiving portion; an absolutely-numbering portion for converting the relative number data and group numbe

Abstract: A division code is added to both ends of a group of pieces of data registered in an item of a data record L and between each pair of adjacent data, so that a piece of code-added data is produced. Thereafter, a character chain pattern (C1,C2) of each pair of adjacent characters C1 and C2 in the code-added data is prepared, an occurrence frequency Fi of each character Ci in each piece of data H is calculated, and character chain information (F1,F2,H,L) corresponding to each character chain pattern (C1,C2) is prepared. Therefore, a index file in which pieces of character chain information prepared for the same item of a plurality of data records are listed is prepared. When a keyword is input by a user, the division code is added to both ends of the keyword to produce a code-added keyword, and a series of keyword character chain patterns is prepared from the code-added keyword in the same manner.

Abstract: A plurality of characters arranged in a text are numbered in order of occurrence for each character type, and an occurrence number table in which one or more occurrence numbers are listed for each character type is prepared in an occurrence number preparing unit. Also, a set of two character types is defined as a character pattern, and a set of two occurrence numbers of two adjacent characters of which types correspond to one character pattern is defined as an occurrence number pattern. One or more occurrence number patterns respectively corresponding to one character pattern are arranged for each character pattern, and a list of occurrence number patterns is prepared as an index file in an index file preparing unit.