JPH04225471A - Keyword retrieving method - Google Patents

Abstract

PURPOSE:To improve the retrieval efficiency by relieving rationally misreading without eliminating it simply, although the retrieval efficiency drops by that portion, when there is misreading in a part of a keyword constitution character. CONSTITUTION:In each character of the same rank as a read character-string and a registered keyword, that which coincides above a first prescribed number corresponding to the number of registered keyword characters is selected, and with regard to each selected read character-string thereof, a character set of every rank is derived. In its character set of every rank, when a character which coincides with a character of the corresponding rank of the registered keyword exists above a second prescribed number corresponding to the number of characters belonging to each character set, all of the read character-strings selected previously are regarded as keywords to be retrieved.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] This invention is a method of searching a document using the same character string as a registered keyword as a keyword, and in particular, even if some of the keyword constituent characters are misread, the method does not simply eliminate them. This invention relates to a keyword search method that can improve search efficiency by providing reasonable relief.

0002

2. Description of the Related Art Generally, it is effective to use keywords to quickly and accurately understand the contents of a document. For example, documents related to global environmental protection issues use keywords such as "radiation,""ozonelayer," and "global pollution." Now, when converting a document into a database, the characters of the document, that is, the original characters, are sequentially read by a character reading device, and the character with the highest degree of similarity to the standard character is selected as the reading character, and a text expressed by a character code is created. This text is searched for the same single character or character string as the registered keyword as a keyword.

0003

In the conventional method, if there is an error in the result read by a character reading device, that is, even one character is misread, the keyword is excluded from the search target even though the keyword exists. That is, when search efficiency is defined as the ratio of the number of search keywords to the total number of keywords in a document, the search efficiency decreases significantly.

[0004] An object of the present invention is to solve the above-mentioned problems of the conventional technology, and even if there is a misreading of some of the characters constituting a keyword, it can be remedied rationally without simply eliminating it. The object of the present invention is to provide a keyword search method that can improve efficiency.

[0005]

[Means for Solving the Problem] In order to solve this problem, the keyword search method according to claim 1 is a method of searching a document using the same character string as a registered keyword as a keyword. read by a character reading device, select a read character string that has the same number of characters as the registered keyword, has a first predetermined number or more of characters in the same rank, and does not match all of the characters of the registered keyword; For each of the selected reading character strings, a set of characters for each same rank is determined; out of the set of characters for each rank, the number of characters that match the characters of the corresponding rank of the registered keyword is a second predetermined number. If there are any of the above, all of the read character strings are considered to be keywords to be searched.

[0006] In the keyword search method according to claim 2, in the method according to claim 1, the first predetermined number is determined according to the number of characters of the registered keyword.

The keyword search method according to claim 3 is the method according to claim 1 or 2, in which the second predetermined number is:
It is determined according to the number of common characters belonging to the character set for each rank.

[0008]

[Operation] In the keyword search method according to claim 1, (1) Each character of the document is read by a character reading device, and if the characters having the same number of characters as the registered keyword and having the same rank match each other at least a first predetermined number, the characters are registered. Select read character strings that do not match all characters of the keyword, that is, perform a first selection as salvageable candidates for read character strings that partially match the registered keyword and would normally be excluded. Find a set of characters for each same rank for each read character string, and when the number of characters that match the characters of the corresponding rank of the registered keyword among the character sets for each rank is equal to or greater than a second predetermined number; As a secondary selection, all of the read character strings are considered as keywords to be searched. The first predetermined number in the first selection is determined according to the number of characters of the registered keyword as in claim 2, and the second predetermined number in the second selection is determined according to the character set for each rank as in claim 3. Each is determined according to the number of common characters it belongs to.

[0009]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A search device to which the keyword search method according to the present invention is applied will be described below with reference to the drawings. FIG. 3 is an illustrative diagram of registered keywords and first selection reading character strings related to the search device. In Figure 3, registered keyword K
is "Fuji Electric", which is composed of four characters. Assume that five read character strings W1 to W5 are selected as a result of the first selection. Namely, they are Fudojoki, Fudodenki, Fueshi Raiki, Gyushidanki, and Soradodenki. Each read character string has a registered keyword and two identical underlined characters. Here, the first predetermined number in the invention is two.

FIG. 4 is a selection process diagram relating to the secondary selection of read character strings. In Figure 4, the first column shows the character ranking, the second column shows the registered keyword, the third column shows the set of each ranking character of the first selection read character string, and the fourth column shows the constituent characters of the registered keyword and each ranking character. The number of matching characters with the set is shown respectively. For example, the constituent characters of the registered keyword corresponding to character rank 1 are "tomi", and the set of characters corresponding to character rank 1 in the five read character strings W1 to W5 that were selected in the first round is {tomi, Wealth, flute, palace, space}. In other words, the number of matching characters is 2. Here, the second predetermined number in the invention is two. Similarly, a character set is obtained for each character rank 2, 3, and 4, and the number of matching characters 2, 2, and 4 is obtained. Therefore, 1
The five read character strings W1 to W5 selected for the next selection also pass the second selection and are determined to be keywords. The first and second predetermined numbers are basically determined empirically,
Generally, the first predetermined number takes a larger value as the number of characters constituting the registered keyword increases, and the second predetermined number takes a proportionally larger value as the number of characters belonging to the character set of each rank increases.

FIG. 2 is a block diagram showing the configuration of the search device. In FIG. 2, reference numeral 1 denotes an image scanner that obtains images related to original characters of a document, and 2 a reading unit that performs reading based on images related to original characters. Reference numeral 3 denotes a memory for read characters, in which read characters are stored as character codes. 4 is an input section for registered keywords, and 5 is a memory for registered keywords. Reference numeral 6 is a collation unit that performs the above-mentioned primary and secondary selection processes, such as collating the corresponding character codes from each memory 3 and 5 and determining whether they match or do not match, and finally considers them as keywords. Decide whether or not. 7 is CR
Press T to display the matching results on the screen. In addition, this CRT7
A printer can be installed to print and output the verification results.

FIG. 1 is a flowchart showing the operation of the search device. In step S1 of FIG. 1, read character strings W having the same number of characters as the keyword KW are sequentially selected (preliminary selection). In step S2, in the read character string W selected in S1, K
A read character string Wi in which all characters in the same rank as W are the same is selected. This is a conventional search method. Step S3
Then, from the read character strings not selected in S2, read character strings Wj having the same characters at least A (first predetermined number) in the same rank as KW are selected (first selection). That is, steps after step S3 are search keyword relief measures. Note that in the example of FIG. 3, A=2. In step 4, steps S1 to S3 are repeated for the entire document. In step S5, B (second predetermined number) or more of the set of characters of the same rank related to the read character string Wj selected in S3 are KW.
Judge whether it is the same as the same ranking character (secondary selection), and select YE.
If S, the process moves to step S5, and if NO, the process ends without relief. Note that in the example of FIG. 4, B=2. Step S6
Then, the read character string Wj is regarded as KW and is rescued. Therefore, the search keywords are the read character strings Wi, Wj.

As described above, the read character strings excluded by the conventional method are rescued through the primary and secondary selection processes, but this rescue is performed under appropriate and reasonable constraints. Therefore, the search efficiency can be improved by the rescue, and the search accuracy of the rescued read character string is high.

[0014]

Effects of the Invention In the keyword search method according to any one of claims 1 to 3, characters having the same number of characters and the same rank as the registered keyword match each other at least a first predetermined number, and each of the registered keywords The read character strings that do not match all the characters are selected as the first selection; for each selected read character string, the set of characters for each same rank is determined, and the correspondence of registered keywords among the character sets for each rank is determined. When the number of characters that match the characters in the ranking is greater than or equal to a second predetermined number, all of the read character strings are considered as keywords to be searched as a secondary selection.

[0015] Therefore, even if some of the characters constituting the keyword are misread, the search efficiency can be improved by rationally repairing the characters rather than simply eliminating them. In particular, as in claim 2, the first predetermined number is determined according to the number of registered keyword characters, and as in claim 3, the second predetermined number is determined according to the number of common characters belonging to the character set for each rank. Therefore, the rescue is performed under appropriate and reasonable constraints, and the search accuracy of the rescued read character string is high.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing the operation of a search device to which the method according to the present invention is applied.

[Figure 2] Block diagram showing the configuration of this search device

[Figure 3]
An illustration of registered keywords and first selection read character strings related to this search device

[Figure 4] Selection process diagram for secondary selection of read character strings

[Explanation of symbols]

1 Image sensor 2 Reading section 3. Memory 4 Input section 5. Memory 6　　　　　　Verification section 7 CRT

Claims (3)

[Claims] Claim 1: In a method of searching a document using the same character string as a registered keyword as a keyword, each character of the document is read by a character reading device, and characters having the same number of characters as the registered keyword and of the same rank are identified. 1st
Select a read character string that matches a predetermined number or more and does not match all of the characters of the registered keyword; Find a set of characters for each same rank for each selected read character string; When the number of characters in the character set that match the characters of the corresponding rank of the registered keyword is equal to or greater than a second predetermined number, all of the read character strings are regarded as keywords to be searched. Keyword search method. 2. The method according to claim 1, wherein the first predetermined number is determined according to the number of characters of the registered keyword. 3. The method according to claim 1 or 2,
A keyword search method characterized in that the second predetermined number is determined according to the number of common characters belonging to a character set for each rank.