JPH07262200A - Statement sorter - Google Patents

Abstract

PURPOSE:To perform classification more desirably by employing such constitution that a key word table registered in advance by making correspond to a classification number is prepared and the classification number of a coincident key word is classified by retrieving the table sequentially. CONSTITUTION:An input device 1 inputs a character string desired to classify. An input character managing means 2 manages a character inputted from the input device 1. A key word matching means 3 takes out the classification number when a part or all of the key words in the key word table 4 coincide, and classifies an inputted character string to a taken out classification number. In this way, detail classification can be performed by reducing the number of key words by performing classification by the combination of the key word, or more desirable classification in which a relating key word with priority is conformed first can be performed by setting a priority classification order.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sentence classification device for classifying input sentences.

[0002]

2. Description of the Related Art Conventionally, a keyword character string matching method has been used for categorizing, storing, and databaseizing news. According to the matching method of the keyword character strings, the keywords are classified, for example, "car".

[0003]

Conventionally, as described above, a simple word, for example, the above-mentioned "car" is extracted as a keyword character string. Therefore, there is a problem that it is not possible to classify Japanese sentences having opposite meanings such as "fast car" and "slow car", or "fast car" and "not fast car".

Further, in the past, for example, "fast car",
In order to classify "fast-running car" and "fast-running car" into the same "fast car", there is also a problem that each must have a keyword string.

The present invention solves these problems.
One or more keywords associated with the classification number,
And prepare a keyword table in which priority classification order is registered in advance, classify them into matching keyword classification numbers by sequentially searching them, and classify with one or a combination of multiple keywords to reduce the number of keywords for fine classification. It is intended to realize the above or to further preferentially realize the classification to a desired classification destination by setting the priority classification order.

[0006]

FIG. 1 shows a block diagram of the principle of the present invention. In FIG. 1, the input device 1 is for inputting a character string to be classified.

The input character management means 2 manages the characters input from the input device 1. The keyword matching means 3 extracts the classification number when a part or all of the keywords in the keyword table 4 match the input character string.

The keyword table 4 is a table in which keywords and priority classification orders are registered in advance in association with classification numbers. The classification result storage unit 5 stores the classification number classified by the keyword matching unit 3 in the sentence storage table 6 in association with the input character string.

The sentence storage table 6 stores the classification result in association with the input character string.

[0010]

The present invention responds to the input of a character string to be classified from the input device 1, as shown in FIG.
With respect to the input character string, the keyword matching means 3 extracts a classification number when a part or all of the keywords in the keyword table 4 match, and classifies the input character string into this extracted classification number. .

With respect to the input character string, the keyword matching means 3 extracts the classification number when all or some of the plurality of keywords in the keyword table 4 match, and extracts the input character string. I try to classify them into classification numbers.

Further, with respect to the character string input by the keyword matching means 3, a part or all of the keyword of the classification number set in the priority classification order when one or all of the keywords in the keyword table 4 match When the numbers match with each other, the classification number is extracted. On the other hand, when the numbers do not match, the classification number matching one keyword is extracted, and the input character string is classified into the extracted classification number.

Further, with respect to the character text input by the keyword matching means 3, when a part or all of a plurality of keywords in the keyword table 4 are matched, a keyword of a classification number set in a priority classification order and a part or all of the keywords. When the numbers match, the classification number is extracted. On the other hand, when the numbers do not match, the classification number that matches a plurality of keywords is extracted, and the input character string is classified into the extracted classification number.

In these cases, when a part or all of the character string input by the keyword matching means 3 matches one or a plurality of keywords, the one having a larger number of matching keywords and the one having a larger number of keywords are When the keywords are the same, the keyword has a larger number of characters, when the keywords have the same number of characters, the longest keyword is the same, and when the keywords have the same length, the keywords have a larger total number of characters.

Further, the keyword is a character string in which kana, kanji, farfabet, and numbers are mixed. Therefore, a keyword table 4 in which one or a plurality of keywords and priority classification order are registered in advance in association with the classification numbers is prepared, and these are sequentially searched to be classified into the classification numbers of the matching keywords, or further classification priority is given. By categorizing in order, the number of keywords can be reduced by one or a combination of a plurality of keywords to realize fine categorization, and the priority categorization order can be set in the keyword table 4 so that the related priority keywords are It is possible to perform a more desirable classification by matching first.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the construction and operation of an embodiment of the present invention will be described in detail with reference to FIGS.

The operation of the configuration of FIG. 1 when the keyword table 4 (part 1) of FIG. 3 is used will be described in detail according to the order shown in the flowchart (part 1) of FIG. Figure 2
In step S1, an input request is made to the user.

In S2, the user inputs a sentence to be classified (character string, input sentence) in response to the input request in S1. For example, enter the following “This problem was solved by the genetic algorithm method” described on the right side.

In step S3, it is determined whether or not the process has ended. If YES, the process ends (END). If NO, the process proceeds to S4. In S4, the input sentence is matched with the keyword table 4 of FIG. 3A or 3B.

In step S5, it is determined whether or not it has been found. this is,
By the matching in S4, it is determined whether or not a part or the whole of the input sentence is found that matches (matches) any of the keywords in the keyword table 4 of FIG. 3A or 3B. In the case of YES, the matched classification number is taken out from the keyword table 4 in S6 and added to the input sentence, and in S7, the classification number, the keyword and the input sentence are displayed in association with each other as a result display. on the other hand,
If NO, the process returns to S3. For example, in the case of the above-mentioned input sentence “This problem was solved by the genetic algorithm method”, matching is performed with the keyword table 4 in FIG. 3A, and here, the keyword “genetic algorithm” is used.
Since it matches with (matches), the classification number 3 at this time
Since 0 is obtained, as a result display, classification number: 30 keyword: "genetic algorithm" input sentence: "this problem was solved by the genetic algorithm method" is displayed.

As described above, in response to the user inputting a sentence and instructing the classification, whether the input sentence partially or wholly matches the keyword table 4 of FIG. 3A or 3B. Matching is performed, a classification number that partially or wholly matches is extracted, the classification number is added to the input sentence and stored, and the classification number, the keyword, and the input sentence are displayed in association with each other. As a result, it is possible to classify any character string such as Chinese characters, kana, katakana, alphabets, numbers, etc. registered in the keyword table 4 in advance and determine the classification number.

FIG. 3 shows an example (1) of the keyword table of the present invention. This is an example of the keyword table 4 used in the flowchart of FIG. FIG. 3 (a)
Shows an example of a keyword table. In the illustrated keyword table 4, keywords are registered in advance in association with classification numbers. Here, kanji, hiragana,
It consists of numbers and alphabets.

FIG. 3B shows an example of the keyword table. The keyword table 4 shown in the figure is a table in which keywords are registered in advance in association with classification numbers, and here, kanji, hiragana, numbers, alphabets, symbols and the like are used.

FIG. 4 shows an example of the sentence storage table of the present invention. In this sentence storage table 6, the keyword matching means 3 matches the input sentence with the keyword in the keyword table 4 of FIG. 3A or 3B, and assigns a classification number to the matched portion (underlined portion). Is a schematic representation. In this way, when a character string is input, the matching part when the keyword matches a preset keyword in the keyword table 4 and its classification number are displayed in association with each other, thereby automatically classifying the input sentence. Is possible.

Next, the operation of the configuration of FIG. 1 when the keyword table 4 (part 2) of FIG. 6 is used will be described in detail according to the order shown in the flowchart (part 2) of FIG. In FIG. 5, S11 requests an input from the user.

In S12, in response to the input request in S11,
The user inputs a sentence to be classified. For example, enter the following "Fast car" on the right side.

In step S13, it is determined whether or not the process has ended. If YES, the process ends (END). If NO, S14
Proceed to. In S14, matching with the keyword table 4 in FIG. 6 is performed to determine whether all the keywords are included in the input sentence. Since a plurality of keywords are registered in advance in association with the classification numbers in the keyword table 4, matching is performed here whether all the keywords are included in the input sentence (in the case of logical product). In addition to this, in the case of logical sum, the classification number is taken out when any of the keywords matches. Here, only the case of the logical product will be described below.

In step S15, it is determined whether or not it has been found. This determines whether or not a match with all keywords in the keyword table 4 of FIG. 6 was found in the input sentence by the matching in S14. If YES, in S16, the matching classification number is extracted from the keyword table 4 and added to the input sentence, and in S7, the classification number, the keyword, and the input sentence are displayed in association with each other as a result display. On the other hand, if NO, the process returns to S13. For example, in the case of the above-mentioned input sentence “fast car”, matching is performed with the keyword table 4 in FIG. 6, and here, since it matches with keyword 1 “fast” and keyword 2 “car”, the classification number 1 at this time is
Since 0 is obtained, as a result display, • Classification number: 10 • Keyword 1: “fast”, keyword 2: “car” • Input sentence: “fast car” is displayed.

As described above, in response to the user inputting a sentence and instructing classification, a plurality of keywords in the keyword table 4 in FIG. The matching classification number is taken out, this classification number is added to the input sentence and stored, and the classification number, the keyword and the input sentence are displayed in association with each other. With these, when the logical product of a plurality of keywords such as kanji, kana, katakana, alphabets and numbers registered in the keyword table 4 in advance is satisfied (there is also a case where the logical sum is satisfied in some cases), the classification is performed. It is possible to determine the classification number.

FIG. 6 shows an example (2) of the keyword table of the present invention. This is an example of the keyword table 4 used in the flowchart of FIG. 5, in which a plurality of keywords are registered in association with the classification numbers, and here, when all the keywords are included in the input sentence, they match. It was done.

Next, the operation of the configuration of FIG. 1 when the keyword table 4 (part 3) of FIG. 8 is used will be described in detail according to the order shown in the flowchart (part 3) of FIG. In FIG. 7, in S21, an input request is made to the user.

In S22, in response to the input request in S21,
The user inputs a sentence to be classified. For example, enter the following "male" on the right side.

In step S23, it is determined whether the processing is finished. If YES, the process ends (END). If NO, S24
Proceed to. In S24, it is matched with the keyword table 4 of FIG. 8 whether the input sentence includes a keyword. This is because the keyword is registered in advance in association with the classification number in the keyword table 4, so matching is performed as to whether this keyword is included in the input sentence.

In step S25, it is determined whether or not it has been found. This determines whether or not a match with the keyword in the keyword table 4 in FIG. 8 was found in the input sentence by the matching in S24. If YES, the process proceeds to S26. N
If it is O, the process returns to S23.

In step S26, it is determined whether the priority classification order is set for the matching keywords in the keyword table 4. If YES, the process proceeds to S27. In the case of NO, since the priority classification order has not been set, the classification number of the matched keyword is determined in S30, and the process proceeds to S31.

In step S27, it is determined whether the priority classification order is over. In the case of YES, since the priority classification order has not been set, the classification number of the matched keyword is determined in S30, and the process proceeds to S31. On the other hand, if NO, the process proceeds to S28.

In step S28, the classification numbers set in the priority classification order are fetched, and it is matched whether or not the keyword of the fetched classification number (priority classification number) is in the input sentence. In the case of YES, since it matches with the keyword of the classification number in the priority classification order, the matching classification number is determined in S29, and the process proceeds to S31. On the other hand, if NO, the process returns to S27.

In step S31, the classification number is extracted from the keyword table 4 and added to the input sentence. For example, as described on the right side, "classification number: 1100" is added to the input sentence "male".

In S32, the result is displayed. For example, in the case of the above-mentioned input sentence “male”, the keyword table 4 in FIG.
Since it matches the key word “male” with the classification number 1000 here, the priority classification order is “120”.
“0” and “1100” are sequentially taken out, and the keywords “male” and “gender” of the classification number are matched with the input sentence. Here, since they match the keyword “male”, the classification number 1100 is determined, The results are displayed as: -Classification number: 1100-Keyword: "male" -Input sentence: "male"

As described above, in response to the user inputting a sentence and instructing the classification, the input sentence is matched or matched with the keyword in the keyword table 4 of FIG. The keyword of the classification number is preferentially matched, and the matching classification number is preferentially displayed in association with the input sentence. Accordingly, by setting the keywords and the classification numbers in the classification priority order in the keyword table 4 in advance, it becomes possible to determine the classification numbers in the priority order for the characters included in the input sentence.

FIG. 8 shows an example (No. 3) of the keyword table of the present invention. This is an example of the keyword table 4 used in the flowchart of FIG. 7, in which the keyword and the classification priority order are registered in association with the classification number, and when the input sentence matches the keyword, the classification is performed with priority. It is possible to arbitrarily specify the number (keyword).

Next, the operation of the configuration of FIG. 1 when the keyword table 4 (part 4) of FIG. 10 is used will be described in detail according to the order shown in the flowchart (part 4) of FIG. In FIG. 9, S41 requests an input to the user.

In S42, in response to the input request in S41,
The user inputs a sentence to be classified. For example, enter the following "soccer ball" on the right side.

S43 is an input sentence and keyword table 4
Matches with until the end. This is because the keyword is registered in advance in association with the classification number in the keyword table 4 of FIG. 10, so matching of whether this keyword is included in the input sentence is performed until the end.

In step S44, it is determined whether it has been found. This determines whether or not a match with the keyword in the keyword table 4 in FIG. 10 was found in the input sentence by the matching in S43. If YES, the process proceeds to S45.
If NO, the process ends (END).

In S45, it is determined whether there is one matching. If YES, the classification number is determined in S46, and the process proceeds to S48. On the other hand, in the case of NO, since a plurality of keywords match, the classification number having the longest number of characters of the keyword is determined in S47, and the process proceeds to S48.

In S48, the classification number is extracted from the keyword 4 and added to the input sentence. For example, as described on the right side, "classification number: 100" is added to the input sentence "sucker ball".

In step S49, the result is displayed. For example, in the case of the input sentence “sucker ball” described above, matching is performed with the keyword table 4 of FIG.
It was determined that the classification number of "soccer ball", which has a long number of characters in the keyword, matches with the keyword "soccer" of 0 and the keyword "soccer ball" of classification number 100, but the result is ・ Category number: 100 ・ Keyword: "Soccer ball" ・ Input sentence: "Soccer ball" is displayed.

As described above, in response to the user inputting a sentence and instructing the classification, all matching is performed to determine whether or not the input sentence includes a keyword that matches the keyword in the keyword table 4 in FIG. The category number of the longest keyword among the keywords is determined and displayed in association with the input sentence. As a result, it becomes possible to register each keyword with respect to the character string that is classified as different depending on the character string to which the keyword is added in advance in the keyword table 4, and classify into the longest keyword classification number.

FIG. 10 shows an example (No. 4) of the keyword table of the present invention. This is an example of the keyword table 4 used in the flowchart of FIG. 9, in which the keywords are registered in association with the classification numbers, all the input sentences that match the keywords are extracted, and the one with the longest keyword is selected. It becomes possible to classify into classification numbers.
For example, the classification number keyword 40 soccer 100 soccer ball shown in the figure is registered, and when a part of the input sentence matches these keywords, it is determined as the classification number of the longest matching keyword.

Next, the flowchart of FIG. 11 (part 5)
The operation of the configuration of FIG. 1 when the keyword table 4 (No. 5) of FIG. 12 is used will be described in detail in the order shown in FIG.

In FIG. 11, in S51, an input request is made to the user. In S52, the user inputs a sentence to be classified in response to the input request of S51. For example, enter the following "Scratch" on the right side.

In step S53, it is determined whether or not the process has ended. In the case of YES, the matching with all the keywords in the keyword table 4 has been completed, so the process is ended (END). On the other hand, if NO, the process proceeds to S54.

In step S54, matching is performed as to whether all keywords are included in the input sentence. This is performed by matching, for example, whether all of the keywords in the keyword table 4 of FIG. 12 are included in the input sentence. here,
It is found that the input sentence “scratch” includes two keywords “scratch” and keyword “scratch” in the classification number 2000, and all of them are included in the input sentence.

In step S55, it is determined whether or not a match has been found. If YES, the process proceeds to S56. If NO, the process returns to S53. A step S56 decides whether or not the priority classification order is set for the matched keywords in the keyword table 4. If YES, the process proceeds to S57. In the case of NO, since the priority classification order has not been set, the classification number of the matched keyword is determined in S60, and the process proceeds to S61.

In step S57, it is determined whether the priority classification order is over. In the case of YES, since the priority classification order has not been set, the classification number of the matched keyword is determined in S60, and the process proceeds to S61. On the other hand, if NO, S
Proceed to 58.

In step S58, the classification numbers set in the priority classification order are extracted, and it is matched whether or not the keyword of the extracted classification number (priority classification number) is in the input sentence. In the case of YES, since it matches with the keyword of the classification number in the priority classification order, the matching classification number is determined in S59, and the process proceeds to S61. On the other hand, if NO, the process returns to S57.

In step S61, the classification number is extracted from the keyword table 4 and added to the input sentence. For example, as shown on the right side, enter "Classification number: 2000" in the sentence "Scratch"
Added to.

In step S62, the result is displayed. For example, in the case of the above-mentioned input sentence “scratch”, it is matched with the keyword table 4 of FIG. 12, and here, since it matches with the keyword “scratch” of the classification number 2000, the priority sorting order is “2.
100 "and" 2200 "are sequentially taken out, and the keywords" scratch "and" easy to attach "of the classification number are matched with the keywords" scratch "and" hard to attach "and the input sentence, and here, the classification number 2100, Since the 2200s do not match, the classification number is determined to be 2000, and the result is displayed as: classification number: 2000-keyword: "scratch", "attach" -input sentence: "scratch"

As described above, in response to the user inputting a sentence and instructing the classification, matching is performed for the keyword in the keyword table 4 in FIG. 11 in the input sentence, and matching is performed, and the matching priority of the matching keyword is determined. The keyword of the classification number is preferentially matched, and the classification number when it matches is preferentially displayed in correspondence with the input sentence.
If they do not match, the original matching classification number is displayed in association with the input sentence. From these, the keyword table 4
By previously setting the keyword and the classification number in the classification priority order, it becomes possible to determine the classification number in the priority order for the characters included in the input sentence.

FIG. 12 shows an example (No. 5) of the keyword table of the present invention. This is an example of the keyword table 4 used in the flowchart of FIG. 11, in which the keywords and the classification priority order are registered in association with the classification numbers, and when the input sentence matches the keywords, the classification is performed with priority. It is possible to arbitrarily specify the classification number (keyword).

Next, the flowchart of FIG. 13 (No. 6)
The operation of the configuration of FIG. 1 when the keyword table 4 (part 6) of FIG. 14 is used will be described in detail in the order shown in FIG.

In FIG. 13, in step S71, an input request is made to the user. In S72, the user inputs a sentence to be classified in response to the input request of S71. For example, enter the following "unisex" on the right side.

In step S73, the matching of all the keywords of the input sentence is performed until the end of the keyword table. This is performed until the end of the keyword table 4 by, for example, matching whether all the keywords 1, 2, 3, etc., which are registered in advance in association with the classification numbers of the keyword table 4 in FIG. .

In step S74, it is determined whether or not the input sentence including all the keywords is found by matching. Yes
In the case of, the matching with all the keywords in the keyword table 4 is completed, and the process is ended (END).
On the other hand, if NO, the process proceeds to S75.

In S75, it is determined whether or not one has been found. If YES, the classification number is determined in S76, and the process proceeds to S78. On the other hand, in the case of NO, since a plurality of matching classification numbers are found, the classification number having the largest number of keywords is determined in S77, and the process proceeds to S78.

In step S78, the classification number is extracted from the keyword table 4 and added to the input sentence. For example, as described on the right side, "classification number: 3000" is added to the input sentence "unisex".

In step S79, the result is displayed. For example, in the case of the above-mentioned input sentence "unisex", matching is performed with the keyword table 4 of FIG. 14, and here, the keyword 1 "male" of the classification number 200, the keyword 2 "female", and the keyword 1 "of the classification number 3000" are used. Man ", keyword 2
Since it matched two of "woman" and keyword 3 "combined use",
The classification number 3000 with the largest number of keywords is determined, and the result is displayed as: -Classification number: 3000-Keywords: "male", "female", "combined" -Input sentence: "unisex"

As described above, in response to the user inputting a sentence and instructing classification, matching is performed for a plurality of keywords in the keyword table 4 of FIG. In some cases, the matching classification number is determined to be the classification number having the larger number of keyword characters, and the input sentence is displayed in association with the determined classification number and keyword. With these, by setting a plurality of keywords in the keyword table 4 in advance,
It is possible to determine the classification number for any combination of individual and entire character strings that can be divided into a plurality of character strings included in the input sentence.

FIG. 13 shows an example (6) of the keyword table of the present invention. This is an example of the keyword table 4 used in the flowchart of FIG. 13, in which a plurality of keywords are registered in association with classification numbers, and when the input sentence matches a plurality of keywords, the number of keywords is the longest. It becomes possible to classify into classification numbers.

[0071]

As described above, according to the present invention,
One or more keywords associated with the classification number,
And a keyword table 4 in which priority classification order is registered in advance
Is prepared, and these are sequentially searched to be classified into matching keyword classification numbers. Therefore, it is possible to classify with one or a combination of a plurality of keywords to reduce the number of keywords and realize fine classification. Further, by setting the priority classification order in the keyword table 4, it is possible to perform a more desirable classification in which the related priority keywords are matched first. With these, (1) For example, the keyword combination “fast AND car” is registered in the keyword table 4 and “fast car”, “fast running car”, and “fast running car” are classified as input sentences. It becomes possible to classify each into a "fast car".

(2) In addition, a combination of two or three keywords "slow AND car" and "fast AND not AND car" is registered in the keyword table 4, and the input sentences related to both are " It can be classified as a "slow car". As a result, it has become possible to perform fine classification even if the number of keywords is reduced.

[Brief description of drawings]

FIG. 1 is a principle block diagram of the present invention.

FIG. 2 is a flowchart (No. 1) for explaining the operation of the present invention.

FIG. 3 is an example (1) of a keyword table of the present invention.

FIG. 4 is an example of a sentence storage table of the present invention.

FIG. 5 is a flowchart (No. 2) for explaining the operation of the present invention.

FIG. 6 is an example (No. 2) of a keyword table of the present invention.

FIG. 7 is a flowchart (No. 3) for explaining the operation of the present invention.

FIG. 8 is a keyword table example (3) of the present invention.

FIG. 9 is a flowchart (No. 4) for explaining the operation of the present invention.

FIG. 10 is an example (4) of the keyword table of the present invention.

FIG. 11 is a flowchart for explaining the operation of the present invention (No. 5).
Is.

FIG. 12 is a keyword table example (5) of the present invention.

FIG. 13 is a flowchart for explaining the operation of the present invention (No. 6).
Is.

FIG. 14 is an example (6) of the keyword table of the present invention.

[Explanation of symbols]

 1: Input device 2: Input character management means 3: Keyword matching means 4: Keyword table 5: Classification result storage means 6: Sentence storage table

Claims (6)

[Claims] 1. A keyword table (4) in which a keyword is registered in association with a classification number, and a classification number when a part or all of the inputted character string matches a keyword in the keyword table (4). Keyword matching means to retrieve (3)
And a sentence classification device that classifies the input character string into the extracted classification number. 2. A keyword table (4) in which a plurality of keywords are registered in association with classification numbers, and an input character string partially or wholly matches all of the plurality of keywords in the keyword table (4). And a keyword matching means (3) for extracting a classification number at that time, and classifying an input character string into the extracted classification number. 3. A keyword table (4) in which one keyword and priority classification order are registered in association with classification numbers.
And when a part or all of the inputted character string matches a part or all of one keyword in the keyword table (4), a part or all of the entered character string matches the keyword of the classification number set in the above priority classification order. And a keyword matching means (3) for extracting the classification number corresponding to one keyword when they do not match, and classifying the input character string into the extracted classification number. A sentence classification device characterized by. 4. A keyword table (4) in which a plurality of keywords and priority classification orders are registered in association with classification numbers.
When a part or all of the input character sentence matches a part or all of a plurality of keywords in the keyword table (4), a part or all of the input character string matches a keyword of a classification number set in the priority classification order. And a keyword matching means (3) for extracting a classification number that matches a plurality of keywords when they do not match, and classifies the input character string into the extracted classification number. A sentence classification device characterized by. 5. When the input character string partially or wholly matches one or more keywords, the number of matching keywords is larger, and the number of keywords is the same when the number of keywords is the same. 5. The method according to claim 1, wherein when the number of characters in the keyword is the same, the longest keyword is classified into the longest keyword, and when the keyword in the longest is the same, the total number of characters of the keyword is classified into the largest. Sentence classification device. 6. The sentence classification device according to claim 1, wherein the keyword is a character string in which kana, kanji, falfabet, and numbers are mixed.