JPH0887501A - Topic level control method and topic structure recognition device in topic structure recognition - Google Patents

Abstract

(57) [Summary] [Purpose] To prevent the topic level in the result of topic structure recognition from becoming deeper than necessary, and to output a topic structure that is easy for humans to understand. [Structure] A topic level increase / decrease table for increasing / decreasing the topic level value according to conditions, and a topic level setting for setting a new topic level regardless of the value when a predetermined condition is satisfied Use a table and. First, the topic level is updated according to the topic level increase / decrease table (step 201), and if the condition of the topic level setting table is satisfied, the topic level is set to a value corresponding to the condition in the topic level setting table. (Step 203).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for topic structure recognition in natural language analysis, and more particularly to a topic level control method for determining a topic level and a topic structure recognition apparatus to which this topic level control method is applied. .

[0002]

2. Description of the Related Art If a human being is presented with text or dialogue data and is given the task of "obtaining the same block written in these texts or dialogue data and the" same thing "", there is no difference between individuals. The property of answering the same structure has been confirmed experimentally. Regarding the experiment, for example, “Takeshita et al .:“ Research on Human Communication from the Perspective of Topic Structure Recognition ”” IEICE 1
Autumn Meeting 993 D-62 (p.6-64) ”. Such a structure understood by humans is called a “topic structure”. Since the topic structure forms a nested structure, each topic has a “topic word” that indicates the topic, a “topic level” that indicates the depth of nesting, and which sentence from which sentence in the text or dialogue data it is. It can be expressed by the "topic scope" of whether to continue until. In the following, the text or dialogue data that is the subject of topic structure analysis is called language data.

FIG. 1 shows an example of a topic structure for language data having contents related to a telecommunication policy. The language data starts from the 0th sentence and continues to at least the 770th sentence. The topic level of the topic having the topic word "communication service" is 1, and the topic scope is the range from the 0th sentence to the 770th sentence. In order to simplify the description, a topic word will be used to refer to the topic, such as “topic of“ communication service ””.

Among the topics of "communication service", there are topics of "new service" and "conventional service" whose topic level is 2. The topics of "new service" are from the 125th sentence to the 431st sentence. Has topic scope up to the sentence,
The topic of "traditional service" is from the 432nd sentence to the 77th sentence.
Has a topic scope of up to 0 sentences. In addition, in the topic of "new service", the child topic "service A"
"Service B" is one of the topics in "Conventional service"
There is a child topic called, and each topic scope is the third
The first sentence to the 431st sentence, the 521st sentence to the 770th sentence.

Recognition of such a topic structure by a computer is called topic structure recognition. Several methods for recognizing the topic structure have been proposed so far. Here, a brief explanation will be given on the topic structure recognition method described in "Takeshita:" Video Retrieval System Using Topic Structure Recognition "Information Processing Society of Japan Information Media Workshop 94-IM-15-1". FIG. 2 is a block diagram showing the configuration of an example of a topic structure recognition device used in this recognition method, FIG. 3 is a flow chart showing a topic structure recognition process in this recognition method, and FIG. 4 is a flow chart in this topic structure recognition process. It is a figure showing an example of a flow of processing after topic structure recognition preprocessing.

The conventional topic structure recognition apparatus shown in FIG. 2 has a data input unit 701 for inputting language data, a processing unit 702 for executing various processes, a display unit 703 for displaying the results, a processing result and It is configured by a storage unit 704 that holds data required during the process and a dictionary / rule unit 705 that stores dictionaries and rules used in the topic structure recognition process. The storage unit 704 stores a language data storage unit 710 that stores preprocessed language data, and a topic structure storage unit 71 that holds intermediate processing results and final processing results.
1 and are provided. Further, the topic structure storage unit 711 is provided with a board development storage unit 712, a semantic development storage unit 713, and an integrated topic storage unit 714. On the other hand, dictionaries
The rule unit 705 is provided with a preprocessing dictionary 721, a semantic expansion processing rule 722, a board expansion processing rule 723, and an integrated processing rule 724.

When performing the topic structure recognition processing using this topic structure recognition device, first, as shown in FIG. 3, the topic structure recognition preprocessing 74 for the input language data 730 is performed.
Perform 0. The first step of the topic structure recognition preprocessing 740 is a morphological analysis processing 741 for the input language data 730. In the morphological analysis process 741, the character string of the input language data 730 is divided into words to form a word string, and the part of speech of each word and the inflection of the inflection word are identified. Then, as the second step of the preprocessing 740,
Using the result of morphological analysis as input, simple sentence segmentation processing 742
Perform The single sentence segmentation process 742 is a process of dividing a sentence including a plurality of predicates, such as an embedded sentence or a compound sentence, into a single sentence including only one predicate. As the third step of the preprocessing 740, the prominent noun phrase extraction 743 is executed. The salient noun phrase extraction 743 is a process of inputting the result of the simple sentence segmentation process 742 and extracting the noun phrase most emphasized in each simple sentence. Then, as the fourth step of the preprocessing 740, block recognition 744 is executed. The block recognition 744 is a process of recognizing a block corresponding to a paragraph in text. These processings belonging to the topic structure recognition preprocessing 740 are executed by the processing unit 702 using the preprocessing dictionary 721 in the dictionary / rule unit 705.
The result is the language data storage unit 710 in the storage unit 704.
Stored in.

When the topic structure recognition preprocessing 740 is completed,
The topic development processing is divided into a board development processing 750 and a semantic development processing 760 and executed. Here, the foundation development is
The topic development explicitly indicated by clue phrases such as "first" and "next", chaptering, and bullet points. Semantic expansion is explicit in each topic of basic expansion. It is the development of a topic that is presented and progressed in a way that is not appropriate.

First, as shown in FIG. 3, in the board development processing 750, a topic establishment section determination 751, a topic word determination 752, and a topic scope and topic level determination 753.
These three processes are sequentially performed. Here, the topic establishment section is a section in which a topic is presented and established. In the topic word determination 752, the prominent noun phrase in each topic establishment section is set as a topic word candidate, and the topic word candidate having the highest priority is selected as a topic word. In determining the topic scope and the topic level 753, processing is performed based on the structure such as bullets. The base development processing 750 is a dictionary /
It is executed by the processing unit 702 using the infrastructure expansion processing rule 723 in the rule unit 705, and the result is the infrastructure expansion storage unit 712 included in the topic structure storage unit 711 in the storage unit 704.
Stored in.

A concrete example of the processing in the substrate development processing 750 is shown in FIG. First, the start point of the language data and the vicinity of the clue phrase such as “first” or “next” are determined as the topic establishment section of the base development.
Then, in the topic word determination 752, “communication service” is started from the first topic establishment section, “new service” is started from the second topic establishment section, and “conventional service” is started from the third topic establishment section. , Respectively, have been selected as topic words.

After the base expansion processing 750 is executed, the semantic expansion processing 760 is executed. The semantic expansion processing 760 is similar to the infrastructure expansion processing 750, in that the topic establishment section is determined 76.
1. The process is composed of three processes: 1. topic word determination 762 and topic scope and topic level determination 763. This semantic expansion processing 760 uses the semantic expansion processing rule 722 in the dictionary / rule part 705 and also the basic expansion processing 7
The result of 50 is also used to be executed by the processing unit 702, and the result is stored in the semantic expansion storage unit 713 included in the topic structure storage unit 711 in the storage unit 704.

In the example shown in FIG. 4, paragraphs longer than a certain length are selected as topic establishment sections, and "service A" and "service B" are selected as topic words in them. The topic scope is required from the start point of the above-mentioned topic establishment section to the start point of the next topic establishment section in the infrastructure development. The topic levels are all set to the same level, that is, level 1 in the case of semantic expansion of text.

Finally, an integrated process 770 of the basic expansion and the semantic expansion is performed, and as a result, the topic structure 780 of the entire language data is output. This integration process 770
It is executed by the processing unit 702 by using the topic structure of each of the base expansion process 750 and the semantic expansion process 760 as an input and the integrated process rule 724 in the dictionary / rule unit 705. In the example shown in FIG. 4, as a result of the integration processing, as shown in FIG.
A topic structure 780 similar to that shown in FIG.

In each of the basic expansion and the semantic expansion, the rules (semantic expansion processing rule 722 and basic expansion processing rule 723) for determining the topic establishment section, the topic word, the topic scope, and the topic level are the language data. It depends on the form of communication such as dialogue, monologue, written text, etc. Regarding the differences in topic development styles and topic structure recognition rules depending on the transmission form, and the results of the topic structure recognition experiment, see “Takeshita et al .:“ Study of Human Communication from the Viewpoint of Topic Structure Recognition ”The Institute of Electronics, Information and Communication Engineers 1993 Autumn Meeting D -62 (p.6-64) ”.

[0015]

However, in the above-mentioned conventional topic structure recognition technology, it is easy to recognize the start point of topic nesting, whereas it is necessary to recognize the end point of topic nesting. It is difficult. For this reason, the topic level in the topic structure recognition result tends to become deeper as it goes behind the language data. There is a problem that it becomes difficult and it is treated as a topic level that is significantly different from the topic level recognized by humans. This tendency is particularly noticeable when the language data is long or the topic is diverse.

[0016] As an example, the minutes (minutes) of the assembly
Take up. FIG. 5 is an example of a meeting record (representative question), which is representative of language data including a large number of diverse topics. FIG. 6 shows what the topic structure of the language data shown in FIG. 5 is recognized by the above-mentioned conventional method, and the result is output in the form of a table of contents. Note that, unlike the normal table of contents, in the one shown here, not only the chapter title but also the character strings before and after the place where the chapter title appears in the language data are also shown. Even in the example shown in Fig. 6, the nesting of chapters is deeper toward the latter half of the table of contents, and in the last "5.11.1.11.1.2.2.3.3.10 PKO bill", the nesting of chapters is as many as 10 layers. ing. If the topic level becomes deeper by this much, it will be difficult for humans to understand the relationship between chapters, and the function as a table of contents will decline.

It is an object of the present invention to provide a topic level control method and a topic structure recognition device which can prevent the topic level in the result of the topic structure recognition from becoming unnecessarily deep and output a topic structure that is easy for humans to understand. It is in.

[0018]

A topic level control method of the present invention is a method for determining a topic level in a topic structure recognition process for recognizing a topic structure of language data using a rule prepared in advance. The topic level increase / decrease table that describes the relationship with the increase / decrease and the topic level setting table that describes the condition and the new topic level pair are prepared in advance, and the topic level increase / decrease table and the topic level setting table are used. To decide.

The topic structure recognition device of the present invention uses an input unit for inputting language data, a dictionary / rule unit for storing rules for topic structure recognition, and rules of the dictionary / storage unit. It has a processing unit that performs processing, a storage unit that stores the result of the processing unit, and a display unit that displays the processing result of the processing unit, and the clue phrase is a nested start type, a topic conversion type, or a nested end type. Classes are classified into three types, and the dictionary / rules section registers the word and part-of-speech table in which the word and the part-of-speech are registered for the nest-start-type clue-phrase, and the word and part-of-speech for the topic conversion-type clue phrase. The topic conversion type clue phrase table, the nesting end type clue phrase table in which the word and the part of speech of the nest end type clue phrase are registered, and information on the type of the clue phrase are included. The topic level increase / decrease table describing the relationship between the condition and the increase / decrease of the topic level, and the topic level setting table describing the combination of the condition and the value of the topic level, the storage unit including the language input from the input unit. The language data storage unit includes a language data storage unit that stores information about data and a topic structure storage unit that stores information about a topic structure, and the language data storage unit stores information about a character string and a part of speech of each word included in the language data. Word information table,
A simple sentence information table that stores information about words, clue phrases, and types of clue phrases included in each simple sentence of the language data, and a utterance number information table that stores information including a speaker of each utterance number and a starting simple sentence number. In addition, the topic structure storage unit includes a table that stores information including a topic establishment section that is a range in which topics are presented and established, topic words, topic levels, and topic scopes.

[0020]

According to the present invention, by using the topic level increase / decrease table in which increase / decrease in topic level is described and the topic level setting table in which a pair of condition and new topic level value is described,
It is possible to prevent the topic level of the topic structure obtained as a recognition result from becoming deep. Thereby, for example, when the topic structure is used as the table of contents, the user can more easily understand the flow of the story in the language data.

[0021]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 7 is a block diagram showing the configuration of a topic structure recognition device according to an embodiment of the present invention. This topic structure recognition device is different from the conventional topic structure recognition device shown in FIG.
In particular, the structures of the tables included in the dictionary / rule section 105 are different.

[Structure of Topic Structure Apparatus of this Embodiment] The topic structure recognition apparatus of this embodiment has a data input unit 101 for inputting language data and a processing unit 102 for executing various processes.
The display unit 103 for displaying the result, the storage unit 104 for holding the processing result and the data required during the process, and the dictionary / rule unit 105 for storing the dictionaries and rules used in the topic structure recognition processing. It is configured. Storage unit 104
A language data storage unit 110 that stores preprocessed language data, and a topic structure storage unit 111 that holds intermediate processing results and final processing results are provided. The language data storage unit 110 is provided with a simple sentence information table 115, a word information table 116, and an utterance number information table 117, and the topic structure storage unit 111 includes a board expansion storage unit 112 and a semantic expansion storage unit 113. Integrated topic storage 1
14 are provided. On the other hand, the dictionary / rule unit 105 includes a preprocessing dictionary 121, a semantic expansion processing rule 122, a board expansion processing rule 123, an integrated processing rule 124, a topic level increase / decrease table 125, and a topic level setting table 12.
6, a nesting start type clue phrase table 127, a topic conversion type clue phrase table 128, and a nesting end type clue phrase table 129 are provided.

When the topic structure of language data is analyzed using this topic structure recognition device, the processing is performed in the same manner as the conventional processing flow shown in FIG. The method of determining is different. When using this topic structure recognition device, the topic level increase / decrease table 125 and the topic level setting table 12
The topic level is determined using 6. Although details of these tables 125 and 126 will be described later, the topic level increase / decrease table 125 describes how much the topic level is relatively changed when a certain condition is satisfied, and the topic level setting table is It describes which value (absolute change) the topic level is forcibly changed when a specific condition is satisfied. Then, the clue phrase tables 127 to 129 are used to determine the conditions described in these tables 125 and 126. Hereinafter, a procedure for analyzing a topic structure by the topic structure recognition apparatus according to the present embodiment will be described centering on a topic level determination method.

[Overall Flow of Determining Topic Level in Infrastructure Deployment] The flowchart of FIG. 8 shows a processing procedure for determining a topic level in infrastructure deployment. First, the topic level increase / decrease table 1 for the topic establishment section of the infrastructure development
25, update the topic level (step 2
01). Then, it is judged whether or not the condition of the topic level setting table 126 is satisfied (step 202), and if not satisfied, the topic level determination process is ended as it is, and if satisfied, the topic level setting is performed. Update the topic level according to table 126 (step 2
03), and the topic level determination processing ends.

[Structure of Each Table] FIG. 9 is a diagram showing a structural example of the topic level increase / decrease table 125. In this embodiment, the topic level is increased / decreased according to the transition pattern of the clue phrase type from the immediately preceding clue phrase to the current clue phrase, and the topic level increase / decrease table 125 shows the relationship between the transition pattern and the increase / decrease amount of the topic level. It describes. In this embodiment, the clue phrases such as "first" and "next" are classified into three types, "nesting start type", "topic conversion type", and "nesting end type". Examples of clue phrases that belong to each type are shown in Table 1.

[0026]

[Table 1] As shown in FIG. 9, if the current clue phrase is a nested start type, the topic level is increased by 1 regardless of the type of the immediately preceding clue phrase. If the current clue phrase is the topic conversion type or the nest end type, and the immediately preceding clue phrase is either the nest start type or the topic conversion type, the topic level is not changed. If the current clue phrase is the topic conversion type or the nest end type, and the immediately preceding clue phrase is the nest end type, the topic level is decreased by one.

FIGS. 10A to 10C show nesting start type clue phrase table 127, topic conversion type clue phrase table 128 and nesting end type clue phrase table 1, respectively.
It is a figure which shows the structure of 29, and the value in a figure is a value in the process of processing with respect to the language data example shown in FIG. In order to store the three kinds of clue phrases shown in Table 1, the dictionary / rule section 105 stores these clue phrase tables 127 to 129.
It is included. Each clue phrase table 127-129
Is provided with a field for a clue phrase number, a field for describing a clue phrase word, and a field for describing a part of speech of each word. For example, in the nesting start-type clue phrase table 127, the clue phrase with the clue phrase number 0 is composed of one word “first”, and its part of speech is an adverb.

On the other hand, as described above, the language data storage unit 110 is provided with the simple sentence information table 115 and the word information table 116. These tables 115,1
16 configuration examples are shown in FIGS. 11A and 11B, respectively. The simple sentence information table 116 is for describing information about a simple sentence, which is a unit having only one predicate, and describes the field of each simple sentence number and the clue phrase included in each simple sentence by the clue phrase number. Field and the
A field that describes the type of the clue phrase, a field that represents the utterance number to which the simple sentence belongs (indicates which speaker the simple sentence is from), and a field that represents the start and end word numbers of the simple sentence ( (Word range field) and. The simple sentence number indicates the order of the simple sentence in the language data as a continuous number starting from 0. The value "-1" in the clue phrase and clue phrase type fields indicates that the clue phrase does not exist. Word information table 1
Reference numeral 16 includes a word number field indicating the order of the word in the language data, a field for describing the word, and a field for describing information such as the part of speech of the word. It is set. As described above, the morpheme analysis and the simple sentence segmentation process are performed in the topic structure recognition preprocessing (see FIG. 3), and the results of these processes are the simple sentence information table 115 and the word information table 11.
6 will be stored. The language data storage unit 1
10, the utterance number information table 117 is also provided. The utterance number information table 117 has a utterance number field indicating the order of utterances of the speaker, a speaker field in which the speaker name is described, With the starting simple sentence field that indicates from which simple sentence number the simple sentence of the speaker starts,
It is configured.

To detect a clue phrase in each simple sentence, the value of the word range of the simple sentence information table 115 is extracted and included in this range on the word information table 116 in each of the three types of clue phrase tables 127 to 129. This is done by checking whether there is an item to be displayed. For example, since the word range of the simple sentence number 0 is the word numbers 0 to 12, it is determined whether the items included in the clue phrase tables 127 to 129 are included in the range of the word numbers 0 to 12 of the word information table 116. Find out. In case of simple sentence number 0,
Since no clue phrase is included, the field values of the clue phrase and the clue phrase type associated with simple sentence number 0 are both -1. If a clue phrase is detected, the clue phrase number is recorded in the clue phrase field and its type is described in the clue phrase type field.

FIG. 12 shows a topic level setting table 126.
The example of composition of is shown. Topic level setting table 126
Is composed of a condition field and a field describing a new topic level. When any of the conditions described in the condition field is satisfied, the topic level is set to the value corresponding to the satisfied condition (the value described in the column on the right side of the satisfied condition). While the topic level transition by the topic level increase / decrease table 125 described above is an increase / decrease based on the topic level up to that point, the topic level update by the topic level setting table 126 is the same as the topic level value up to that point. The absolute value of the topic level is specified regardless of. In the example shown in FIG. 12, when the topic level exceeds 5, the topic level is returned to 1. If the speaker change occurs, the new utterance number is composed of 3 or more sentences, and there are 5 or more sentences before the present speaker change, the topic level is returned to 1. If two or more conditions are satisfied at the same time, for example, the one described higher in the topic level setting table 126 may be adopted.

Next, the relationship between each of the infrastructure expansion storage unit 112, the simple sentence information table 115 and the utterance number information table 117 and the conditions described in the topic level setting table 126 will be described with reference to FIG. The values in the figure are values in the process of processing the language data example shown in FIG.

The infrastructure expansion storage unit 112 stores, for each topic number, a field (topic establishment section field) describing a single sentence number indicating the start and end of a topic establishment section in which a topic is presented / established and a topic word. It includes a field for describing, a field for describing a topic level, and a field for describing a topic scope by word numbers of its start and end. As described above, the utterance number information table 11
7 includes a utterance number field, a field for describing the speaker, and a field for describing a simple sentence number at which the utterance number starts. In the illustrated example, the speaker with the utterance number 0 is the "chairperson". Yes, it starts from simple sentence 0 and ends at simple sentence 1 which is one before the starting simple sentence of the next utterance number 1.

Topic level setting table 1 shown in FIG.
Whether "topic level>5" in the condition of 26 is satisfied can be determined by checking the value of the topic level field of the basic expansion storage unit 112. In addition, whether or not the condition “when a speaker change occurs”, which is a condition of the topic level setting table 126, is satisfied is determined by the simple sentence information table 115 for the start simple sentence of the topic establishment section recorded in the infrastructure development storage unit 112. It can be determined by examining the utterance number field of. Further, whether or not the conditions of the topic level setting table 126, "the new utterance number contains three or more simple sentences" and "there are five or more simple sentences included from the start of the language data to the speaker change", are satisfied. Can be determined from the value of the starting simple sentence field of the utterance number information table 117.

[Explanation Using Example of Language Data] Next, referring to FIG.
The topic level control method in the infrastructure development in this embodiment will be described in detail with reference to the example shown in FIG. In the following description, a simple sentence having a simple sentence number N will be referred to as a simple sentence #N for simplicity.

The topic level of the topic number 45 is already stored in the infrastructure development storage unit 112, and its value is 4.
Further, the topic establishment section of the topic number 45 starts from simple sentence # 720. According to the simple sentence information table 115, the simple sentence # 720 belongs to the utterance number 4.

Since the topic level of the topic number 46 has not been decided yet (it is blank in the drawing), it is decided according to the present invention. First, the topic level increase / decrease table 125
According to, update the topic level. The topic establishment section of the topic number 46 is simple sentences # 822 to # 826,
Looking at the range of the simple sentence information table 115, simple sentence # 8
There is a nesting-type clue phrase at 22. As shown in FIG. 9, according to the topic level increase / decrease table 125, if the current clue phrase is a nested start type, the topic level is incremented by 1 regardless of the previous clue phrase.
The topic level is 5.

Next, it is checked whether the conditions of the topic level setting table 126 are satisfied. "Topic level>
The condition "5" is not satisfied because the topic level is 5. The topic establishment section of topic number 46 is simple sentence # 82.
Although it starts from 2, according to the simple sentence information table 115,
This simple sentence belongs to utterance number 6. As described above, since the topic number 45 belongs to the utterance number 4, it can be seen that the speaker change has occurred. Further, according to the utterance number information table 117, the utterance number 6 starts from the simple sentence # 785, and the simple sentence # 10 immediately before the starting simple sentence # 1047 of the next utterance number 7 is # 10.
Since it continues up to 46, the condition that "the new utterance number contains three or more simple sentences" is also satisfied. Further, since the utterance number 6 starts from the simple sentence # 785, the condition that "5 or more simple sentences are included from the start of the language data to the speaker change" is also satisfied. From these, it can be seen that the second condition of the topic level setting table 126 shown in FIG. 11 is satisfied, so that the topic level of the topic number 46 finally becomes 1.

FIG. 14 shows the result of topic structure recognition performed on the language data illustrated in FIG. 5 using the topic level control method shown in this embodiment. In the table of contents shown in FIG. 14, the nesting of chapters is at most quadruple, and the nesting level is returned at a reasonable place. Therefore, when compared with the table of contents of FIG. It is easy.

[0039]

As described above, according to the present invention, by using the topic level increase / decrease table and the topic level setting table, it is possible to prevent the topic level in the topic structure recognition result from becoming deep. There is.
It is very effective for long language data such as conference proceedings and lecture records, and is particularly effective for language data such as conference proceedings in which one utterance number contains many sentences. The present invention makes the recognized word structure easy for the user to understand.

[Brief description of drawings]

FIG. 1 is an example of topic structure recognition by a human.

FIG. 2 is a block diagram showing a structure of an example of a conventional topic structure recognition device.

FIG. 3 is a flowchart showing a conventional process for topic structure recognition.

FIG. 4 shows an example after preprocessing in conventional topic structure recognition.

FIG. 5 is a diagram showing an example of language data.

6 is a diagram showing a result of extracting a topic structure by applying a conventional topic structure recognition method to the language data shown in FIG.

FIG. 7 is a block diagram showing the structure of a topic structure recognition device according to an embodiment of the present invention.

8 is a flowchart showing a topic level determination process in infrastructure development by the apparatus of FIG.

FIG. 9 is a diagram showing a configuration example of a topic level increase / decrease table.

10A to 10C are diagrams showing configuration examples of a nest start type clue phrase table, a topic conversion type clue phrase table, and a nest end type clue phrase table, respectively.

11A and 11B are diagrams showing configuration examples of a simple sentence information table and a word information table, respectively.

FIG. 12 is a diagram showing a configuration example of a topic level setting table.

FIG. 13 is a diagram illustrating the relationship between each of the infrastructure expansion storage unit, the simple sentence information table, and the utterance number information table and the conditions described in the topic level setting table.

14 is a diagram showing an example of a topic structure recognition result performed by applying the method of the present invention using the apparatus of FIG.

[Explanation of symbols]

 101 data input unit 102 processing unit 103 display unit 104 storage unit 105 dictionary / rule unit 110 language data storage unit 111 topic structure storage unit 112 basic expansion storage unit 113 semantic expansion storage unit 114 integrated topic storage unit 115 simple sentence information table 116 words Information table 117 Utterance number information table 121 Preprocessing dictionary 122 Semantic expansion processing rule 123 Infrastructure expansion processing rule 124 Integrated processing rule 125 Topic level increase / decrease table 126 Topic level setting table 127 Nesting start type clue phrase table 128 Topic conversion type clue phrase Table 129 Nesting type clue phrase table 201 to 203 steps

Claims (10)

[Claims] 1. A topic level increase / decrease table for describing a relationship between a condition and an increase / decrease in a topic level in a method of determining a topic level in a topic structure recognition process for recognizing a topic structure of language data using a prepared rule. , A topic level setting table for describing a pair of a condition and a new topic level is prepared in advance, and a topic level is determined using the topic level increase / decrease table and the topic level setting table. 2. The topic structure recognition processing divides the language data into simple sentences that are units having only one predicate, extracts a clue phrase from each of the simple sentences, and identifies the type of the clue phrase. Perform topic structure recognition preprocessing that includes the topic expansion, and then separate the topic expansion into the explicitly indicated basic expansion and the semantic expansion that expands within the basic expansion, and the basic expansion processing rule and the semantic expansion processing rule, respectively. Using, and with respect to the infrastructure development, determination of a topic establishment section in which a topic is presented and established, determination of a topic word in the topic establishment section, and determination of a topic level and topic scope indicating nesting of topics are sequentially performed. , Next, with respect to the semantic expansion, the topic establishment section, the topic word, the topic scope and the topic level are sequentially determined, and thereafter, the basic development and the meaning are determined using the integrated processing rule. A process for performing a process of integration with respect to each of the processing result of the expansion, the topic level control method according to claim 1. 3. The topic level is updated according to the topic level increase / decrease table, then it is determined whether or not the condition of the topic level setting table is satisfied, and if the condition is satisfied, the topic level is determined. The topic level control method according to claim 1, wherein the topic level is updated according to a setting table. 4. The topic level control method according to claim 3, wherein the condition of the topic level setting table includes information on speaker replacement. 5. The topic level control method according to claim 3, wherein the condition of the topic level setting table includes information about the value of the topic level updated by the topic level increase / decrease table. 6. An input unit for inputting language data,
A dictionary / rules section that stores rules for topic structure recognition,
It has a processing unit that performs processing using the rules of the dictionary / storage unit, a storage unit that stores the result of the processing unit, and a display unit that displays the processing result of the processing unit, and the clue phrase starts nesting. Type, topic conversion type, and nest ending type, and the dictionary / rule part registers a nest start type clue phrase table in which the word and part of speech of a nest start type clue phrase are registered, and a topic conversion type For the clue phrase of, the topic conversion type clue phrase table in which the word and the part of speech are registered,
A nesting end type cue phrase table in which the word and the part of speech are registered for the nesting end type clue phrase, and a topic level increase / decrease table describing the relationship between the condition including the information of the type of the clue phrase and the increase / decrease in the topic level, And a topic level setting table that describes a set of conditions and topic level values, wherein the storage unit stores a language data storage unit that stores information about language data input from the input unit and information about a topic structure. Including a topic structure storage unit, the language data storage unit, a word information table storing information about character strings and parts of speech of each word included in the language data, words included in each single sentence of the language data, and a clue A simple sentence information table that stores information about phrases and clue phrase types, and information that includes the speaker and starting simple sentence number of each utterance number The topic structure storage unit includes a table for storing information including a topic establishment section that is a range in which a topic is presented and established, a topic word, a topic level, and a topic scope. Characterized topic structure recognition device. 7. The dictionary / rule part divides the language data into simple sentences which are units having only one predicate, extracts a clue phrase from each of the simple sentences, and identifies the type of the clue phrase. Preprocessing dictionary for topic structure recognition preprocessing including, basic expansion processing rules for processing basic expansion, semantic expansion processing rules for processing semantic expansion, basic expansion and semantic An integrated processing rule for integrating expansion, wherein the topic structure storage unit stores a basic expansion storage unit that stores information about basic expansion, a semantic expansion storage unit that stores information about semantic expansion, a basic expansion and a meaning Including an integrated topic storage unit that stores information after integration of dynamic expansion, each of the semantic expansion storage unit and the base expansion storage unit is a topic establishment section in which a topic is presented and established, and a topic word, Comprising a table for storing information including a title level, a topic scope, the topic structure recognition apparatus according to claim 6. 8. The processing unit updates the topic level according to the topic level increase / decrease table, saves the result in the infrastructure development storage unit, and then whether the conditions of the topic level setting table are satisfied. The topic structure recognizing device according to claim 6, wherein if satisfied, the topic level is updated according to the topic level setting table and the result is stored in the infrastructure expansion storage unit. . 9. The condition of the topic level setting table includes:
The topic structure recognizing device according to claim 8, further comprising a condition for using information about a start simple sentence described in the utterance number information table and an utterance number described in the simple sentence information table. 10. The topic structure recognition device according to claim 8, wherein the condition of the topic level setting table includes a condition of using a value of the topic level stored in the infrastructure expansion storage unit.