JPS62282364A - Character string retrieval system - Google Patents

Abstract

PURPOSE:To attain the retrieval a declinable word with a simple operation by generating all declensions of the declinable word automatically and retrieving them. CONSTITUTION:A character string input means 1 inputs a character string to be retrieved and a retrieval kind designation means 2 designates the retrieval of declinable word expansion. A declinable word recognition means 7 uses a declinable word stem storage means 5 and a declinable word ending storage means 6 to recognize the stem and endings of the declinable word given as the character string to be retrieved. A declension expanding means 8 reads all declension endings from the means 6 and adds the endings to the stem to generate all the declensions, which are written in a retrieval character storage means 9. A character string storage means 10 reads the character string from the retrieval character string storage means 9 to retrieve a sentence stored in a sentence storage means 4 and to find out the character position in the character string to be retrieved, and an edition control means 11 revises the character string to be retrieved displayed on a display means into the display method different from that of other character strings by means of a method such as inverted display.

Description

[Detailed Description of the Invention] 3. Detailed Description of the Invention (Field of Industrial Application) The present invention is the first in a Japanese word processor, etc.
This relates to a character string search method in which a second character string is searched for a first character string. The column search method is popular as a search function in word processors.

In this case, the first character string is a sentence created with a word processor, and the second character string is usually input frequently by the user of the word processor when the search function is specified.

Below, NEC Japanese word processor NWP-1ON'
The search function of Bungo 1 will be explained as an example.

If you press the ``Character Search'' key while a sentence created with a word processor is displayed on the screen, the following will be displayed on the screen (■ represents a cursor).

(Character string search) [■] Here, enter the character string you are looking for in [ ]. For example, to search for the character string ``say'', enter the following:

(Character string search) [Say ■] When you press the "Execute" key after entering this character string to search for (searched character string), the text will be searched from the position where the cursor was before pressing the "Character search" key. At the back, the first character string to be searched that appears in the sentence is searched for, and if it is found, its position is highlighted. At this point, pressing the "Tag" key will continue the search for more sentences after that position. The search ends when you press the "Execute" key instead of the "Tag" key.

(Problems to be Solved by the Invention) The search function in this word processor is mainly used to search for and locate a portion to be processed in some way, such as creation, editing, or reference.

Therefore, the searched character string is usually a word rather than a meaningless character or character string.

However, conventional search functions such as those described above simply match character strings and do not match words. Therefore, when using a word without a conjugation such as a noun as a search string, simple character string matching is often successful, but when trying to search for a conjugated word in a single sentence, It is necessary to repeatedly search for all possible conjugations of the word.

For example, if you are trying to search for the verb "say" in a sentence, use the following character string as the searched character string.
You have to repeat the search ◇ “Koto” (form following i = 7 no j, etc.) “Koto”
(Unnatural form B = the form that follows ``u・yo'') ``I say'' (conjugated form A = the form that follows ``ta'', etc.) ``say'' (conjugated form B = the form that follows ``ta'', etc.)
Final form, adjunctive form) ``Ie'' (hypothetical form, imperative form) In this case, there is a method of using ``word'', the root word of ``say'', as the search string, but in this case, ``word'', ``language'', etc. Many more unnecessary words including ``words'' such as ``utterance'' are searched.

On the other hand, it is technically possible to use a large-scale word dictionary to analyze a sentence to be searched and to recognize all the words in the sentence (Nagaya, Tsujii, Yamagami, Takebe: ``Japanese language "Dictionary Memory and Automatic Segmentation of Japanese Sentences", Magazine "Information Processing" No. 1
9, No. 6, 1978).

Therefore, it is expected that if such an analysis is performed first and then a word is searched based on the analysis result, the word can be searched accurately.

However, the word dictionary used for this analysis is generally 3
The drawback is that it is a large project, ranging from 10,000 to 300,000 words, and requires a huge amount of processing, such as checking sentences against word dictionaries and testing grammatical connections between parts of speech.

SUMMARY OF THE INVENTION An object of the present invention is to provide a character string search method that eliminates these drawbacks and allows searches even for conjugated terms with simple operations.

(Means for Solving the Problems) The character string search method of the present invention targets a first character string input in advance, and searches for a plurality of characters related to a second character string input for searching. a search specifying means for specifying one of the search types; a word stem storage means for storing the word stem of a word and the type of conjugation of the word; a conjugated ending storage means that stores the term in association with the conjugated ending, and a term that identifies the term in the second character string and the conjugated form of the term by referring to the language stem memory unit and the conjugated ending storage unit; equating means, and conjugation expansion means for generating all conjugated forms of the predicate recognized by the term recognition means with reference to the conjugation ending storage means, in one search specified by the search designation means. On the other hand, it is characterized in that all inflections of the predicate generated by the inflection expansion means are searched from the first character string.

(Example) Next, the present invention will be explained in detail using the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of a word processor using the character string search method of the present invention. In the figure, 1 character string input means 1 is for inputting a character string of a Japanese sentence, a keyboard, a kana-kanji conversion input device,
A voice recognition device, a character recognition device, etc. can be used. This character string input means 1 also inputs editing commands such as insertion, deletion, movement, and copying, and also inputs a searched character string (second character string).

The search type designating means 2ri specifies a plurality of search types, and can be realized by providing a plurality of keys or dials corresponding to the search types. Search types include, for example, ``-scattered string search,'' which searches for matching parts as conventional character strings, and ``word expansion search,'' which searches for all conjugations of a predicate. . The display means 3ri is for displaying text strings, text messages, message messages, etc. from the word processor to the user, and uses a CRT display or the like.

The text storage means 4 stores the character string of the text inputted and edited by the character string input means 1, and the IC
A memory, a magnetic disk device, etc. can be used.

The character string stored in the text storage means 4 becomes the search target (first character string) of the character string search method of the present invention.

The word stem storage means 5 is for storing the word stem of a word and the type of conjugation of the word, and may be a device such as an IC memory, a magnetic disk device, or the like.

Table 1 shows an example of the contents of the language stem storage means 5.

Table 1 In this table, the word stem storage means 5 stores the word stem 5 of the word.
0 and the type of conjugation 51 corresponding to this stem 50 are memorized.
It is sufficient to memorize it in the form of a character code written as "Kane"), and then decide on the type of conjugation 51ri, the code corresponding to that type (for example, ``Suyu Shita Ichidan Utilization''), and memorize that code. Bye.

The conjugation ending storage means 6 stores the conjugation endings of the words in association with the types of conjugations, and may be an IC memory, a magnetic disk device, or the like. Table 2 shows an example of the contents of this conjugated word ending storage means 6. Table 2 shows the types of conjugations of the predicates, such as 60ri ``ga-gyo 5-dan'' conjugations, ``su-gyo 5-dan'' conjugations, etc., and conjugation endings 61 to 68 corresponding to the types of conjugations are shown. 61 is unnatural form A (``
62 is the unnatural form B (ru/you).
(form following), 63 is the continuous form (form followed by continuous form), and 63 is the continuous form.

64 is the adjunctive form B (the form that follows r-ta, etc.), 65 is the final form, 66 is the adnominal form, 67 is the hypothetical form, 68 is the imperative form,
The character code of the notation of each conjugated word ending is stored.

> The usage type 60 in Table 2 corresponds to the usage type 51 in Table 1, and a code corresponding to the usage type is stored.

The predicate recognition means 7 refers to the used language stem storage means 5 and the conjugated word ending storage means 6, and determines the predicate in the searched character string input from the character string input means 1 and the conjugated form of the predicate. certify. Specifically, first, the searched character string is
1 to the word recognition means 7, and the word recognition means 7 sends the word stem of the word that matches the substring of the searched character string to the word stem 50 of the word stored in the word stem storage means 5. The conjugation ending storage means 6) obtains the conjugation ending for the type of conjugation for the stem of the searched term. If one of the conjugated endings matches the character string immediately after the stem of the term in the searched character string, the conjugated form corresponding to that conjugated ending becomes the conjugated form of the term. If there are multiple candidates, it is possible to identify the candidates by comparing the conjugated endings only by matching the stem of the word. Finally, the term recognition means 7 sends the stem of the searched term and the type of use of the term to the use development means 8.

The conjugation development means 8 refers to the conjugation ending storage means 6 and generates all conjugations of the predicates recognized by the predicate recognition means 7. Specifically, the conjugation development means 8 receives the stem of the word and the type of conjugation of the word from the conjugation recognition means 7, and extracts all words corresponding to the type of conjugation from the conjugation ending storage means 6. Take out the conjugated ending υ. Then, these conjugated endings are combined with the stem of the word to generate all the conjugated forms of the word. The inflection development means 8 writes all the generated inflections into the search character string storage means 9.

The search character string storage means 9 stores all conjugations of a predicate generated by the conjugation expansion means 8 during the "phrase expansion search." Further, in the case of "-dispersed character string search", the character string to be searched inputted by the character string input means 1 is sent from the editing control means 11 and stored in the search character string storage means 9. As this storage means 9, an IC memory, a magnetic disk device, etc. can be used.

The character string comparison means 10 searches for the character string stored in the search character string storage means 9 from the character strings of the sentences stored in the sentence storage means 4, and returns the found position to the editing control means 11.

Editing control means 111''l: writes the character string of the sentence input from the character string input means 1 into the sentence storage means 4, and writes it into the sentence storage means 4 according to the editing command manually input from the character string input means 1. Rewrites the memorized string of sentences.Furthermore, controls the operation of each component mentioned above,
The character string search method of the present invention is realized.

Next, how the editing control means 11 controls each component of the double series of "-scattered character string search" and "expanded word search" will be explained using an example. In this example, it is assumed that the following sentences are stored in the sentence storage means 4.

``He said that monopoles exist. However, there are many scholars who deny what he said.'' ``-Dispersed string search'' is explained. This is a conventional character string search method.Character string manual means 1゜display means 32 text storage means 4
．． Only the search character string recording means 92, the character string comparison means io, and the editing control means 11 operate.

First, a character string to be searched is input through the character string input means 1,
The search type specifying means 2 specifies "-scattered character string search." In this example, it is assumed that "say" is input as the character string to be searched. The information from the character string input means 1 and the search type designation means 2 is sent to the editing control means 11, and since "-dispersed character string search" has been specified, the editing control means 11 selects the input search character. Search column as string storage means 9!
Get into it.

Subsequently, the editing control means 11 activates the character string comparison means io. The character string comparison means 10 is a search character string storage means 9.
``say'' is read as a character string to be searched, and it is retrieved from the sentences stored in the sentence storage means 4.

As a result, the searched character string "say" is found from the 25th character to the 26th character of the sentence. The character string comparison means 1° then sends the character position (25-26) where the searched character string is found to the editing control means 11. Editing control means 1
1ri, the text stored in the text storage means 4 is displayed on the display means 3, but upon receiving information from the character string comparison means 10, two of the character strings displayed on the display means 3 are displayed.
Change the display method for the 5th to 26th characters to be different from other character strings, such as by displaying them in reverse.

In this way, if you perform a "-scattered character string search" on "say", the 25th to 26th character "say" will be searched, but the 15th to 16th character "say" will be searched. ” will not be searched. If you want to intentionally avoid searching for ``Koi'', you can use the ``-scattered string search'' like I just described, but if you want ``Koi'' to be searched as well, use the first All you have to do is perform a ``term expansion search'' as described below.

Now, "term expansion search" which is a feature of the present invention will be explained.

First, a character string to be searched is input through the character string input means 1,
Search type specifying means 2) ``Term expansion search j is specified. In this example, it is assumed that ri and ``say'' are input as the searched character string. The information from the character string input means 1 and the search type designation means 2 is sent to the 1 editing control means 11, and since the false collection control means 11ri specifies "term expansion search", the input searched character string ``say'' is sent to the predicate recognition means 7 (it is not written to the search string storage means 9).

The term recognition means 7 operates when the searched character string is sent from the editing control means 11. The term recognition means 7d searches the term stem storage means 5 for the stem of the term that matches the partial character string of the sent character string to be searched ``say''. There is no word stem that matches ``say'', but a match is found for ``go'' with the - character omitted (see Table 1). Therefore, the conjugation recognition means 7 retrieves the conjugation type ``wa-gyo 5-dan conjugation'' corresponding to the ``word'' from the conjugation stem storage means 5, and stores it in the conjugation-suffix storage means 6. Read out the conjugated ending for "Godan conjugation". As a result, the following conjugated endings are read out (see Table 2).

``rwa''``o''``i''``ri''``u''``u''``e''``e'' And among these, the string excluding the stem of the term searched from the searched string is `` Search for a match with "U" (or a substring if this is two or more characters). Here, the final form and the adjunctive form "U" match the character string obtained by removing the stem of the word from the searched character string. Therefore, the term recognition means 7 determines that the term given as the searched character string is ``
It is recognized as the final form or conjunctive form of the ``wa-gyo five-dan conjugation'' with ``word'' as the root word. Then, the word recognition means 7 sends "word" and "Wa-gyo 5-dan conjugation" to the conjugation development means 8.

The usage development means 8 operates when information is sent from the predicate recognition means 7. The conjugation development means 8 first reads out all conjugation endings for "wa-gyo 5-dan conjugation" from the conjugation ending storage means 6. The read conjugated ending is the same as the one read by the predicate recognition means 7. Then, the conjugation development means 8 adds these conjugation endings to the stem word ``word'' of the term sent from the predicate recognition means 7 to generate the following full conjugation form.

"Say""Say""Say""Say""Say""Say"
"Ie""Ie" Next, the conjugation development means 8 writes all these conjugations into the search character string storage means 9. However, in this case, identical character strings will be deleted. Therefore, the following six character strings are written into the search character string storage means 9.

"Say""Sayo""Say""Say""Say""Say" Thereafter, the utilization development means 8 notifies the editing control means 11 that the writing of the search target character string is completed.

In response, the editing control means 11 controls the character string comparison means 1
Start 0. The character string comparison means 10 reads the above six character strings from the search character string storage means 9 and searches for them from the aforementioned sentences stored in the sentence storage means 4. As a result, "say" is found in the 15th to 16th characters of the sentence, and "say" is found in the 25th to 26th characters. Then, the character string comparison means 10 sends the character position (15-16') where the character string is found to the editing control means 11.
and send (25-26). The editing control means 41 displays the text stored in the text storage means 4 on the display means 3, and upon receiving information from the character string comparison means 10,
1 from the 15th character of the character string displayed on display means 3
Change the display method of the 6th character and the 25th to 26th characters to be different from other character strings, such as by displaying them in reverse.

(Effects of the Invention) As explained above, according to the character string search method of the present invention, it is possible to perform a search in which all conjugated forms of a term are automatically generated. Conventionally, when doing something similar to this, it was necessary for the user to think of all the conjugations of the word and repeatedly search for strings for each conjugation, but with the present invention, this is done in a single operation. This can be achieved with

In addition, using a word dictionary to identify the words in all sentences and then extracting the words requires a word dictionary containing tens of thousands of words to several dozen words, resulting in an enormous amount of processing. If there are only words such as those used in inventions, only a few hundred to 10,000 words are needed, and the processing efficiency is also good.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment of a word processor using the character string search method of the present invention. 1...Character string input means, 2...Search type designation means, 3...Display means, 4...
Sentence storage means, 5... Language stem storage means, 6.
. . . Conjugation ending storage means, 7. Term recognition means, 8. conjugation development means, 9. Search character string storage means, 10. ...String comparison means,
1.

Claims (1)

[Claims] A search specification means for specifying one of a plurality of search types for a second character string input for searching a first character string input in advance; a conjugation stem storage means for storing conjugation types of words; a conjugation suffix storage means for storing conjugation endings of said words in association with said conjugation types; and conjugation stem storage means and conjugation suffix storage means. a term recognition means for identifying a term in the second character string and a conjugated form of the term by referring to the term;
and a conjugation expansion means for generating all conjugations of the term recognized by the term recognition means by referring to the conjugation ending storage means, and for one search specified by the search specification means, the conjugation A character string search method, characterized in that all conjugated forms of the term generated by an expanding means are retrieved from the first character string.