JP3329352B2 - Topic level control method and topic structure recognition device in topic structure recognition - Google Patents

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 the topic level control method is applied. .

[0002]

2. Description of the Related Art When a human being is presented with text or conversation data and given the task of "seeking the same thing in a block in which the same is written from these text or conversation data", there is no individual difference. The ability to answer the same structure has been experimentally confirmed. For the experiment, for example, "Takeshita et al .:" Research on Human Communication from the Viewpoint of Topic Structure Recognition "IEICE 1
99-93 Autumn Meeting D-62 (p.6-64) ”. Such a structure grasped by a human is called a “topic structure”. Since the topic structure forms a nested structure, each topic consists of a "topic word" indicating the topic, a "topic level" indicating the depth of the nesting, and a sentence of the sentence in the text or dialog data. Can be expressed by the "topic scope" of whether to continue. In the following, text and conversation data for which the topic structure is analyzed will be referred to as language data.

FIG. 1 shows an example of a topic structure for language data having contents related to a telecommunications 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 in the range from the 0th sentence to the 770th sentence. For the sake of simplicity, in the following, the topic will be referred to using a topic word, such as "topic of" communication service "".

[0004] Among the topics of "communication services", there are topics of "new service" having a topic level of 2 and "conventional service". The topics of "new service" are from the 125th sentence to the 431rd sentence. Has a topic scope up to the sentence,
The topic of "traditional services" is from the 432rd sentence to the 77th sentence
It has a topic scope of up to 0 sentences. Also, among the topics of "new service", a child topic of "service A"
"Service B" in the topic of "traditional services"
Child topic exists, and each topic scope is 3rd
It is from the 01st sentence to the 431st sentence, the 521st sentence and the 770th sentence.

Recognition of such a topic structure by a computer is called topic structure recognition. Several methods have been proposed for recognizing topic structures. Here, a brief description of the topic structure recognition method described in "Takeshita:" Video Retrieval System Using Topic Structure Recognition ", Information Processing Society of Japan 94-IM-15-1". FIG. 2 is a block diagram showing a configuration of an example of a topic structure recognition device used in this recognition method. FIG. 3 is a flowchart showing a topic structure recognition process in this recognition method. It is a figure showing an example of the flow of processing after topic structure recognition pre-processing.

The conventional topic structure recognition apparatus shown in FIG. 2 includes a data input unit 701 for inputting language data, a processing unit 702 for executing various processes, a display unit 703 for displaying results, a processing result and The storage unit 704 stores data required during processing, and the dictionary / rule unit 705 stores dictionaries and rules used in the topic structure recognition processing. The storage unit 704 includes 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 is provided. Further, the topic structure storage unit 711 includes a base development storage unit 712, a semantic development storage unit 713, and an integrated topic storage unit 714. On the other hand, a dictionary
The rule unit 705 is provided with a preprocessing dictionary 721, a semantic expansion processing rule 722, a base expansion processing rule 723, and an integration processing rule 724.

When a topic structure recognition process is performed using this topic structure recognition apparatus, first, as shown in FIG. 3, a topic structure recognition preprocessing 74 for input language data 730 is performed.
Perform 0. The first step of the topic structure recognition pre-processing 740 is a morphological analysis process 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, the inflected form of the inflected word, and the like are identified. Subsequently, as a second step of the preprocessing 740,
Using the result of the morphological analysis as an input, a single sentence separation process 742
Perform The single sentence delimiting process 742 is a process of dividing a sentence including a plurality of predicates, such as an embedded sentence or a multiple sentence, into a single sentence including only one predicate. As a third step of preprocessing 740, salient noun phrase extraction 743 is executed. The prominent noun phrase extraction 743 is a process of extracting the noun phrase that is most emphasized in each simple sentence by using the result of the simple sentence separation process 742 as an input. Then, as a 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 processes belonging to the topic structure recognition pre-processing 740 are executed by the processing unit 702 using the pre-processing dictionary 721 in the dictionary / rule unit 705,
The result is stored in the language data storage unit 710 in the storage unit 704.
Is stored in

When the topic structure recognition pre-processing 740 is completed,
The topic development process is separated into a base development process 750 and a semantic development process 760 and executed. Here, infrastructure development
Topic development explicitly indicated by clue phrases such as "first" and "next", chapters, bullet points, etc. It is the development of a topic that is presented and progressed in a non-target form.

First, as shown in FIG. 3, in the base development process 750, a topic establishment section determination 751, a topic word determination 752, a topic scope and a topic level determination 753 are performed.
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, prominent noun phrases in each topic establishment section are set as topic word candidates, and the topic word with the highest priority among these topic word candidates is selected. In the topic scope and topic level determination 753, processing is performed based on a structure such as an itemized list. The base development process 750 is performed by
The processing is performed by the processing unit 702 using the base expansion processing rule 723 in the rule unit 705, and the result is stored in the base expansion storage unit 712 included in the topic structure storage unit 711 in the storage unit 704.
Is stored in

FIG. 4 shows a specific example of the processing in the base development processing 750. First, the starting point of the language data and the vicinity of a 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” from the first topic establishment section, “new service” from the second topic establishment section, and “conventional service” from the third topic establishment section. , Respectively, have been selected as topic words.

After the execution of the base development process 750, a semantic development process 760 is performed. The semantic development process 760 is similar to the base development process 750, and is used to determine the topic establishment section 76
1. It is composed of three processes, namely, topic word determination 762, topic scope and topic level determination 763. This semantic expansion processing 760 uses the semantic expansion processing rule 722 in the dictionary / rule unit 705 and the base expansion processing 7
The processing is also executed by the processing unit 702 using the result of 50, and the result is stored in the semantic development 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 these sections. As the topic scope, the range from the start point of the above-described topic establishment section to the start point of the next topic establishment section in the base development is required. The topic levels are all set to the same level, that is, level 1 in the case of the semantic development of the text.

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

In each of the basic development and the semantic development, the rules for determining the topic establishment section, topic word, topic scope, and topic level (semantic development processing rule 722 and base development processing rule 723) are based on language data. It depends on the form of communication, such as dialogue, monologue, or written text. For the differences between topic development styles and topic structure recognition rules depending on the transmission form, and the results of topic structure recognition experiments, see "Takeshita et al .:" Study of Human Communication from the Viewpoint of Topic Structure Recognition "IEICE 1993 Fall Meeting D -62 (p.6-64)].

[0015]

However, in the above-mentioned conventional topic structure recognition technology, it is easy to recognize the starting point of the nesting of topics, but it is difficult to recognize the ending point of the nesting of topics. It is difficult. For this reason, the topic level in the topic structure recognition result tends to be deeper behind the linguistic data, and when trying to use this topic structure recognition result as a table of contents for humans, for example, There is a problem that it becomes difficult, or the topic level is significantly different from the topic level recognized by humans. This tendency is remarkable especially when the language data is long or the topics are various.

As an example, proceedings (minutes) in a parliament
Take up. FIG. 5 shows an example of a meeting record (representative question), which is a representative example of language data including many diverse topics. FIG. 6 shows a result of recognizing a topic structure in the language data shown in FIG. 5 by the above-described conventional method and outputting the result in the form of a table of contents. In addition, unlike the ordinary table of contents, 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 described. In the example shown in Fig. 6, the nesting of chapters is deeper in 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 tenfold. ing. If the topic level becomes so deep, it becomes difficult for humans to understand the relationship between chapters, and the function as a table of contents is weakened.

An object of the present invention is to provide a topic level control method and a topic structure recognizing device capable of suppressing a topic level in a topic structure recognition result from becoming unnecessarily deep and outputting a topic structure that is easy for humans to understand. It is in.

[0018]

Topic level control method of the present invention, in order to solve the problem] is the topic level method of determining the in recognizing the topic structure recognition apparatus topic structure of the language data using the pre-prepared rule, a new topic level Factors to generate,
Factors that continue topic level, factors that end topic level
The story that appeared last time for three kinds of topic level change factors of the child
Topic level change factors and topic level change factors
The topic level increase / decrease table that describes the increase / decrease value of the topic level by the tricks is stored in advance as a rule.
Conditions for changing level values to absolute values and changes
After the topic level setting table describing the correspondence between the value of the absolute topic level allowed to pre-stored as a rule, the
A topic structure recognition device determines a topic level using the topic level increase / decrease table and the topic level setting table.

The topic structure recognition apparatus of the present invention were used: an input unit for inputting language data, and the dictionary-rule section for storing rules such for topic structure recognition, rules such the dictionary specification and regulations section A processing unit for performing processing, a storage unit for storing the results of the processing units, and a display unit for displaying the processing results of the processing units, wherein the clue phrases are nested start type, topic conversion type, and nested end type. It is classified as One of the type, factors before Symbol dictionary and regulations section, to generate a new topic level, the topic Les
The factor of continuing the bell, the factor of ending the topic level 3
For the topic level change factors, the topic level
And the topic level change factor that appeared this time.
The topic level increase / decrease table that describes the increase / decrease value of the topic level according to the
Between the condition for the change and the absolute topic level value after the change
The topic level setting table, the storage unit includes a language data storage unit that stores information about the language data input from the input unit, and a topic structure storage unit that stores information about the topic structure, language data storage unit, simple sentence information table for storing the word information table for storing information about the strings and part of speech of each word included in the language data, a single word included in the single sentence in each sentence unit of the language data The topic structure storage unit includes a table for storing information including a topic establishment section, a topic word, a topic level, and a topic scope, which are ranges in which topics are presented and established.

[0020]

The present invention uses a topic level increase / decrease table describing the increase / decrease in topic level and a topic level setting table describing a pair of a condition and a new topic level value.
It is possible to prevent the topic level of the topic structure obtained as a result of recognition from becoming deep. Thus, 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 the topic structure recognition device according to one embodiment of the present invention. This topic structure recognition device is different from the conventional topic structure recognition device shown in FIG.
In particular, they differ in the configuration of tables included in the dictionary / rule unit 105.

[Configuration of Topic Structure Recognition Apparatus of the Embodiment] The topic structure recognition apparatus of the embodiment includes a data input unit 101 for inputting linguistic data, a processing unit 102 for executing various processes, and displaying a result. Display unit 103, and a storage unit 104 for holding processing results and data required during processing.
And a dictionary / rule unit 105 for storing dictionaries and rules used in the topic structure recognition processing. The storage unit 104 includes a language data storage unit 110 that stores language data after preprocessing, and a topic structure storage unit 111 that holds intermediate processing results and final processing results. The linguistic data storage unit 110 has a simple sentence information table 1
15, word information table 116, and utterance number information table 1
17 is provided, on the topic structure storage unit 111, based on
A board development storage unit 112, a semantic development storage unit 113, and an integrated topic storage unit 114 are provided. On the other hand, the dictionary / rule unit 105 includes a preprocessing dictionary 121, a semantic expansion processing rule 122, a base expansion processing rule 123, and an integrated processing rule 124.
A topic level increase / decrease table 125, a topic level setting table 126, a nested start type clue phrase table 127, a topic conversion type clue phrase table 128, and a nested end type clue phrase table 129.

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

[Overall Flow of Determining Topic Level in Base Development] The flowchart of FIG. 8 shows a processing procedure for determining a topic level in base development. First, the topic level increase / decrease table 1 for the topic establishment section of the base development
25, the topic level is updated (step 2
01). Then, it is determined whether the condition of the topic level setting table 126 is satisfied (step 202). If the condition is not satisfied, the process of determining the topic level is terminated as it is. The topic level is updated according to the table 126 (step 2
03), the topic level determination processing ends.

[Configuration of Each Table] FIG. 9 is a diagram showing a configuration example of the topic level increase / decrease table 125. In the present embodiment, the topic level is increased or decreased by 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 indicates the relationship between this transition pattern and the increase / decrease amount of the topic level. Has been described. In this embodiment, clue phrases such as "first" and "next" are classified into three types: "nested start type", "topic conversion type", and "nested end type". Examples of clue phrases belonging to each type are shown in Table 1.

[0026]

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

FIGS. 10A to 10C show the nested start type clue table 127, the topic conversion type clue table 128, and the nested end type clue table 1 respectively.
FIG. 29 is a diagram showing a configuration of the language data shown in FIG. 29, and the values in the figure are values in the process of processing the example language data shown in FIG. In order to store the three types of clue phrases shown in Table 1, the dictionary / rule unit 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 word of the clue phrase, and a field for describing the 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 (a) and (b) are shown in FIGS. The simple sentence information table 116 is for describing information about a simple sentence which is a unit having only one predicate, and describes a field of each simple sentence number and a clue phrase included in each simple sentence by the clue phrase number. Fields and
A field describing the type of the clue phrase, a field indicating the utterance number to which the simple sentence belongs (indicating the number of the speaker of the simple sentence), and a field indicating the start and end word numbers of the simple sentence ( Word range field). The simple sentence number indicates the number of the simple sentence in the language data by a sequential number starting from 0. The value "-1" in the clue phrase and clue phrase type fields indicates that there is no clue phrase. Word information table 1
16 includes a word number field indicating the number 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. Is set. As described above, the morphological analysis and the simple sentence delimiting process are performed in the topic structure recognition pre-processing (see FIG. 3), and the processing results are stored in the simple sentence information table 115 and the word information table 11.
6 will be stored. The language data storage unit 1
The utterance number information table 117 is also provided in the utterance number table 10. The utterance number information table 117 includes an utterance number field indicating the order of utterance of the speaker, a speaker field in which the speaker name is described, The starting simple sentence field indicating which simple sentence number the simple sentence of the speaker starts with,
It is configured.

For the detection of the clue phrase in each simple sentence, the value of the word range of the simple sentence information table 115 is taken out, and within this range on the word information table 116, the three types of clue phrase tables 127 to 129 are included. This is done by examining whether the item to be found exists. For example, since the word range of simple sentence number 0 is word numbers 0 to 12, it is determined whether items included in clue phrase tables 127 to 129 are included in the range of word numbers 0 to 12 of word information table 116. Find out. For simple sentence number 0,
Since the clue phrase is not included, each of the field values of the clue phrase and the clue phrase type associated with the simple sentence number 0 is set to −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.
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 a value corresponding to the satisfied condition (the value described in the right column of the satisfied condition). The transition of the topic level according to the topic level increase / decrease table 125 described above is an increase / decrease based on the previous topic level, whereas the topic level setting table 126 updates the topic level, and Irrespective of the topic level, the absolute value of the topic level is specified. 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 three or more single sentences, and if there are five or more single sentences before the current 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 the base development storage unit 112, the single 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 example language data shown in FIG.

The base development storage unit 112 stores, for each topic number, a field describing a single sentence number indicating the start and end of a topic establishment section in which a topic is presented / established (topic establishment section field), and a topic word. A field for describing, a field for describing a topic level, and a field for describing a topic scope by its start and end word numbers are included. As described above, the utterance number information table 11
7 includes a speech number field, a field describing the speaker, and a field describing the simple sentence number at which the speech number starts. In the illustrated example, the speaker with the speech number 0 is “chairperson”. Yes, it starts with a single sentence 0, and ends with a single sentence 1 that is immediately before the start single sentence of the next utterance number 1.

Topic level setting table 1 shown in FIG.
Whether “topic level> 5” in the condition 26 is satisfied can be determined by examining the value of the topic level field in the base development storage unit 112. Whether or not “when a speaker change occurs” as a condition of the topic level setting table 126 is satisfied, the single sentence information table 115 is used for the start single sentence of the topic establishment section recorded in the base development storage unit 112. Can be determined by examining the utterance number field of. Whether or not the condition of the topic level setting table 126 satisfies “3 or more simple sentences are included in the new utterance number” and “5 or more simple sentences are included from the start of language data to the change of speaker”. Can be determined from the value of the start simple sentence field of the utterance number information table 117.

[Explanation Using Example of Language Data] Next, FIG.
The topic level control method in the base development in this embodiment will be described in detail using 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 base development storage unit 112, and its value is 4.
The topic establishment section of the topic number 45 starts from a simple sentence # 720. According to the single sentence information table 115, the single sentence # 720 belongs to the utterance number 4.

Since the topic level of the topic number 46 has not yet been determined (blank in the figure), it is determined according to the present invention. First, the topic level increase / decrease table 125
, The topic level is updated. The topic establishment section of topic number 46 is from simple sentences # 822 to # 826,
Looking at the range of the simple sentence information table 115, the simple sentence # 8
22 has a nested start clue phrase. As shown in FIG. 9, according to the topic level increase / decrease table 125, if the current clue phrase is the nesting 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 or not the condition of the topic level setting table 126 is 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
2, but according to the simple sentence information table 115,
This single 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. According to the utterance number information table 117, the utterance number 6 starts from the single sentence # 785, and the single sentence # 10 immediately before the start single sentence # 1047 of the next utterance number 7
Since it continues up to 46, the condition that “3 or more simple sentences are included in the new utterance number” is also satisfied. In addition, 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 change of the speaker” 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 a result obtained by performing topic structure recognition on the language data illustrated in FIG. 5 using the topic level control method described in this embodiment. In the table of contents shown in FIG. 14, the chapter nesting is at most four folds, and the nesting level is returned at a more appropriate point. Therefore, 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 deeper. There is.
This is very effective for long linguistic data such as conference proceedings and lecture records, and is particularly effective for linguistic data such as conference proceedings where one utterance number contains many sentences. According to the present invention, the recognized word structure is easy for the user to understand.

[Brief description of the drawings]

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

FIG. 2 is a block diagram showing the 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 is an example after the pre-processing in the 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. 5;

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

8 is a flowchart showing a topic level determination process in a base development by the apparatus of FIG. 7;

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

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

FIGS. 11 (a) and 11 (b) are diagrams each showing a configuration example of a single sentence information table and a word information table.

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

FIG. 13 is a diagram for explaining the relationship between each of a base development storage unit, a simple sentence information table, and an utterance number information table and conditions described in a topic level setting table.

14 is a diagram showing an example of a topic structure recognition result performed by using the apparatus of FIG. 7 and applying the method of the present invention.

[Explanation of symbols]

 Reference Signs List 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 base development storage unit 113 semantic development storage unit 114 integrated topic storage unit 115 single sentence information table 116 words Information table 117 Speech number information table 121 Preprocessing dictionary 122 Semantic expansion processing rule 123 Base expansion processing rule 124 Integration 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 Nested end type clue table 201-203 steps

Continuation of front page (56) References JP-A-7-160710 (JP, A) JP-A-7-160711 (JP, A) JP-A-7-160712 (JP, A) JP-A-6-236410 (JP) , A) JP-A-6-139276 (JP, A) JP-A-4-332084 (JP, A) Atsushi Takeshita, "Video Retrieval System Using Topic Structure Recognition", Information Processing Society of Japan Research Report 94-IM-15 , Japan, March 11, 1994, Vol. 94, no. 24, p. 1-p. 8. Atsushi Takeshita, et al., "Detection of topic introductions in D-67 monologue" 1994 IEICE Autumn Conference, Japan, September 5, 1994, p. 70 (58) Field surveyed (Int.Cl. ⁷ , DB name) G06F 17/21-17/30

Claims (10)

(57) [Claims] 1. A method for determining a topic level in a topic structure recognition device for recognizing a topic structure of language data using rules prepared in advance, wherein a factor for generating a new topic level and a topic level are continued.
Factors and factors that terminate topic levels
The topic level change factor that appeared last time
Talking about the matrix of topic level change factors that appeared this time
A topic level increase / decrease table that describes the increase / decrease
A condition for changing the value of the topic level to an absolute value,
A topic level setting table describing the correspondence with the absolute topic level value after the change is stored in advance as a rule , and the topic structure recognizing device uses the topic level increase / decrease table and the topic level setting table. A topic level control method characterized by determining a topic level. 2. The topic level is updated according to the topic level increase / decrease table, and it is determined whether a condition of the topic level setting table is satisfied. If the condition is satisfied, the topic level is updated. The topic level control method according to claim 1, wherein the topic level is updated according to a setting table. 3. Language data using rules prepared in advance.
Topics in a topic structure recognizer that recognizes topic structures
Factors that generate new topic levels in how to determine bells, continue topic levels
Factors and factors that terminate topic levels
The topic level change factor that appeared last time
Talking about the matrix of topic level change factors that appeared this time
A topic level increase / decrease table that describes the increase / decrease
A condition for changing the value of the topic level to an absolute value,
A topic that describes the correspondence with the absolute topic level value after the change
Were allowed prestored level setting table as rules, the topic structure recognition apparatus divides the language data in the simple sentence is a unit having only one predicate, including Rukoto issuing extract the clue phrase from each sentence performs topic structure recognition preprocessing, then the topic structure recognition apparatus performs the determination of the topic established interval topics is presented-established, the determination of topic words in the topic established intervals, further extracted clue clause Based on
The topic level change factor is calculated based on the
Determines the talk <br/> title level by using the table and the topic level setting table, further, the determination of topics scope
A topic level control method characterized in that : 4. In the topic structure recognition preprocessing, each unit is
Associating identifiable information speaker sentences, including a turn-taking as a condition for the topic level setting table
No, the speaker associated with the simple sentence
4. The topic level control method according to claim 3, wherein the topic level control method is performed based on information that specifies the topic level. 5. The topic level control method according to claim 3, wherein the condition of the topic level setting table includes information on a topic level value updated by the topic level increase / decrease table. 6. An input unit for inputting language data,
A dictionary / rule part for storing rules for topic structure recognition;
It includes a processing unit that performs processing using the dictionary specification and regulations of the rules include, a storage unit for storing the result of the processing unit, and a display unit for displaying the processing result by the processing unit, pre-Symbol Dictionary and regulations Factors that generate new topic levels
Child, factor to continue topic level, end topic level
Appearing last time for three topic level change factors
Topic level change factors and topic level change factors that appeared this time
And topic level decrease table that describes the variation value for the topic level by a matrix of the values of the topic level to an absolute value
Conditions for change and absolute topic level after change
A topic level setting table that describes the correspondence with the value , wherein the storage unit stores a language data storage unit that stores information about language data input from the input unit, and a topic structure storage unit that stores information about a topic structure. wherein the storing the language data storing unit, and the word information table for storing information about the strings and part of speech of each word included in the language data, a single word included in the single sentence in each sentence unit of the language data to and a single sentence information table, the topic structure storage unit, topic presented, including a table for storing information including the topic establishment section and topic words and the topic level and topic scope ranges to be established, that Topic structure recognition device that features. 7. The processing unit updates a topic level in accordance with the topic level increase / decrease table, and outputs a result of the updating.
The topic level is stored in the topic structure storage unit. Next, it is checked whether the condition of the topic level setting table is satisfied. If the condition is satisfied, the topic level is updated according to the topic level setting table, and the result is updated. The topic structure recognition device according to claim 6, wherein is stored in the topic structure storage unit. 8. A clue phrase is a nested start type and a topic conversion type.
And the nested-end type clue phrase.
Nested cue phrases that register the words and parts of speech
Words about tables and topic-based clue phrases
A topic conversion type clue table that registers
Words and parts of speech for nested clue phrases
Including the registered nested end type clue table
Further, the linguistic data is divided into simple sentences which are units having only one predicate, clue phrases are extracted from each simple sentence, and the topic structure recognition pre-processing includes identifying the type of the clue phrase. includes a pretreatment dictionary, the topic level increase or decrease table, clue phrase appearing last
Type and the clue type matrix that appeared this time
7. An increase / decrease value of a topic level is described in the list.
2. A topic structure recognition device according to claim 1. 9. The utterance information is specified in the simple sentence information table.
Information for the topic level setting table
As a single sentence.
The topic structure recognition device according to claim 8, wherein the recognition is performed based on information specifying the linked speaker . 10. The topic structure recognition apparatus according to claim 8, wherein the condition of the topic level setting table includes a condition using a topic level value stored in the topic structure storage unit.