JP7078244B2 - Data processing equipment, data processing methods, data processing systems and programs - Google Patents

Description

The present invention relates to a data processing device, a data processing method, a data processing system and a program for processing posted text data collected from the Internet. More specifically, the present invention relates to a technique for giving a title, a headline, etc. to a group of posted sentences classified by content.

Many sentences are posted on the Internet every day. In particular, the social networking service (SNS) allows you to easily post comments and photos, so information about incidents, accidents, disasters, etc. may be posted in real time, such as posted photos and posted videos. Is beginning to be used in the news.

On the other hand, in order to extract information that can be a news source from a wide variety of sentences posted on the Internet, it is necessary to read the posted sentences one by one and check the contents, which takes time and effort. In addition, news such as incidents, accidents, and disasters are required to be breaking news, but in order to know the occurrence of these events from posts on the Internet, it is necessary to constantly monitor the posted text, which is inefficient.

Therefore, recently, a service that collects specific posts and articles from the Internet and provides them to users is attracting attention. However, even when using such a service, it is necessary to confirm whether each sentence is "related to the target event" or "reliable". As a method of efficiently judging the contents of sentences created by others in a short time, there is a method of using the titles and headlines attached to them, and conventionally, it is automatically added to posted articles and newspaper articles. A technique for giving a title has been proposed (see Patent Documents 1 and 2).

For example, in the method described in Patent Document 1, a unique expression as a feature word for a title is extracted by using a past article in the same category as the input article, and a title of the input article is created. Further, in the method described in Patent Document 2, a candidate sentence for creating a title sentence is extracted from a newspaper article, and a subject or a predicate necessary for a title sentence for a database is extracted from the candidate sentence to create a title sentence. is doing.

Japanese Unexamined Patent Publication No. 2010-191851 Japanese Unexamined Patent Publication No. 9-44497

However, all of the above-mentioned conventional title creation methods are techniques for giving a title indicating the content to one article, and give a title or a headline to a posted sentence group consisting of a plurality of posted sentences by different contributors. It does not create short sentences such as summaries or articles that show the contents. Therefore, even if the techniques described in Patent Documents 1 and 2 are applied, it is not possible to create a short sentence showing the contents of a group of posted sentences relating to an arbitrary event collected via the Internet.

Therefore, the present invention provides a data processing device, a data processing method, a data processing system, and a program capable of creating a short sentence showing the contents of a group of posted sentences related to an incident, an accident, a disaster, etc. collected via the Internet. With the goal.

The data processing device according to the present invention analyzes each posted sentence of a posted sentence group composed of two or more posted sentences related to the same event collected via the Internet, and appears a word included in the posted sentence group. The sentence analysis unit that ranks by frequency selects words that can identify the event that is common to each posted sentence from the words included in the posted sentence group, and the content of the event is used using the selected words. It has a short sentence creation unit that automatically creates a short sentence representing the above, and an event word database that stores a nomenclature that can identify an event and is used when selecting a word that can identify an event in the short sentence creation unit. The described event is an incident, an accident, or a disaster, and in the selection of words that can identify the event in the short sentence preparation unit, the word having the highest frequency of appearance is preferentially selected.
The sentence analysis unit includes a part-speech decomposition unit that decomposes part of speech for each posted sentence, a noun extraction unit that extracts nouns from words that have been decomposed into part speech, and a noun counting unit that ranks the extracted nouns according to their appearance frequency. In that case, the short sentence creation unit can create the short sentence by using, for example, a noun having a high frequency of appearance.
The short sentence creation unit is selected by, for example, a word selection unit that selects a noun that can identify an event common to each posted sentence of the posted sentence group from among the nouns included in the posted sentence group, and the word selection unit. It is equipped with a short sentence generation unit that creates short sentences using the nouns, and the word selection unit compares the noun data stored in the event word database with the noun ranking data created by the sentence analysis unit. However, it is also possible to preferentially select nouns with a high frequency of appearance.
In that case, the noun that can identify the event can be a noun representing a region and / or a noun representing an event.
On the other hand, the data processing apparatus of the present invention may further have a ranking data storage unit for storing ranking data of the frequency of appearance of the word.
The short sentence is, for example, a title or a headline.

The data processing method according to the present invention is a method of automatically creating a short sentence using one or a plurality of data processing devices, and two or more posts relating to the same event collected via the Internet by the data processing device. A sentence analysis process that analyzes each posted sentence of the posted sentence group composed of sentences and ranks the words included in the posted sentence group according to the frequency of appearance, and each posted sentence from the words included in the posted sentence group. A short sentence creation step of selecting a word that can identify the event common to the above and creating a short sentence expressing the content of the event using the selected word is performed, and the event is an incident, an accident, or a disaster. In the selection of words that can identify the event in the short sentence creation process , a word database in which nomenclature that can specify the event is stored is used, and words with a high frequency of appearance are preferentially selected.
In the data processing method of the present invention, in the sentence analysis step, a step of part-speech decomposition for each posted sentence, a step of extracting nouns from the part-speech-decomposed words, and a step of ranking the extracted nouns by appearance frequency. And may be carried out, and the short sentence may be created by using a noun having a high frequency of appearance in the short sentence creation step.
Further, in the short sentence creation step, a step of selecting a noun that can identify an event common to each posted sentence of the posted sentence group from the nouns included in the posted sentence group, and a short sentence using the selected noun. In that case, for example, in the step of selecting the noun, the noun data stored in the event word database is compared with the noun ranking data created in the sentence analysis step. , Nouns with high frequency of appearance can be selected with priority.

The data processing system according to the present invention analyzes each posted sentence of a posted sentence group composed of two or more posted sentences related to the same event collected via the Internet, and appears a word included in the posted sentence group. From the sentence analysis device that ranks by frequency and the words included in the posted sentence group, the words that can identify the event common to each posted sentence are selected, and the content of the event is used using the selected words. It has a short sentence creation device that automatically creates a short sentence representing the above, and an event word database that stores a nomenclature that can identify an event and is used when selecting a word that can identify an event in the short sentence creation device. The described event is an incident, an accident, or a disaster, and in the selection of a word capable of identifying the event in the short sentence creating device, a word having a high frequency of appearance is preferentially selected.
The data processing system of the present invention further classifies a plurality of posted sentences collected via the Internet for each event, and is a group of posted sentences which is a collection of text data composed of two or more posted sentences related to the same event. The posted sentence classification device may be provided, and in that case, the sentence analysis device analyzes the posted sentence group created by the posted sentence classification device.
Further, the data processing system of the present invention may have a distribution device that adds the short sentence data created by the short sentence creation device to the data of the posted sentence group and distributes it externally.

The program according to the present invention analyzes each posted sentence of a posted sentence group composed of two or more posted sentences related to the same event collected via the Internet on a computer, and obtains words included in the posted sentence group. A sentence analysis function that ranks by frequency of appearance and a word that can identify the event common to each posted sentence are selected from the words included in the posted sentence group, and the selected word is used to identify the event. The short sentence creation function that automatically creates a short sentence representing the content is executed, and the event is an incident, an accident, or a disaster, and the event can be specified by selecting a word that can identify the event in the short sentence creation function. Using the event word database in which nomenclature is stored, words with the highest frequency of appearance are prioritized and selected.

In the present invention, the "posted text group" refers to a collection of text data composed of two or more posted texts relating to an arbitrary event collected via the Internet.
In addition, the "event" in the present invention includes various events such as incidents, accidents, and disasters that are taken up in the media, and the place of occurrence is not limited to Japan but also includes overseas.

According to the present invention, in a posted sentence group collected via the Internet, the frequency of appearance of a specific word is analyzed, and a high-ranking word is extracted to generate a short sentence. Therefore, an event related to the posted sentence group. It is possible to automatically create a short sentence showing the contents of.

It is a conceptual diagram which shows the structural example of the data processing apparatus of 1st Embodiment of this invention. It is a flowchart which shows the method of performing the data processing using the data processing apparatus 1 shown in FIG. It is a figure which shows the example of the rank data. It is a conceptual diagram which shows the structural example of the data processing system of the 2nd Embodiment of this invention.

Hereinafter, embodiments for carrying out the present invention will be described in detail with reference to the accompanying drawings. The present invention is not limited to the embodiments described below.

(First Embodiment)
First, the data processing apparatus according to the first embodiment of the present invention will be described. FIG. 1 is a conceptual diagram showing a configuration example of the data processing device of the present embodiment. As shown in FIG. 1, the data processing device 1 of the present embodiment has a sentence analysis unit 2 and a short sentence creation unit 3, and automatically creates short sentences related to a group of posted sentences such as a title and a headline. The data processing device 1 may be provided with a ranking data storage unit 4, an event word database 5, or the like, if necessary.

[Sentence analysis unit 2]
The sentence analysis unit 2 analyzes each posted sentence of the input posted sentence group, and ranks the words contained in the posted sentence according to the frequency of appearance. Here, the "posted sentence group" analyzed by the sentence analysis unit 2 is a collection of text data composed of two or more posted sentences related to an arbitrary event collected via the Internet. Further, the "posted text" may be anything that can be collected via the Internet, and examples thereof include information posted on SNS, blogs, electronic bulletin boards (Bulletin Board System: BBS), and the like. The length of the posted text to be analyzed is not particularly limited, and may be long or short.

The structure of the sentence analysis unit 2 is not particularly limited, but for example, a part-speech decomposition unit 21 that decomposes part of speech for each posted sentence, a noun extraction unit 22 that extracts nouns from words decomposed by part of speech, and extraction. It can be configured to include a noun counting unit 23 that ranks the nouns that have been given by the frequency of appearance. The sentence analysis unit 2 has a structure in which parts of speech other than nouns are also extracted and ranked, or all words decomposed by part of speech are ranked and only nouns are extracted from the ranked words. You can also do it.

The analysis method of each posted sentence in the sentence analysis unit 2, that is, the processing method in the part of speech decomposition unit 21, the noun extraction unit 22, the noun counting unit 23, and the like is not particularly limited, but for example, text mining and word counting. And methods such as full-text search technology can be applied.

[Short sentence creation section 3]
The short sentence creation unit 3 creates a short sentence related to the posted sentence group based on the ranking data of the words (or nouns) included in the posted sentence group created by the sentence analysis unit 2. Examples of the short sentences created by the short sentence creation unit 3 include those that specify an event such as a title and a headline, as well as summaries and articles that briefly summarize the specific contents of the event.

The method of automatically creating a short sentence in the short sentence creation unit 3 is not particularly limited, but for example, a noun representing a specific content such as a region or an event is selected from high-ranked words (or nouns) and selected. It is possible to use a method of creating a sentence using the nouns that have been given. In that case, the short sentence creation unit 3 includes a word selection unit 31 that selects a noun representing a specific content from the nouns included in each posted sentence of the posted sentence group based on the ranking data of the word (or noun), and a word. A short sentence generation unit 32 that creates a short sentence using the nouns selected by the selection unit 31 may be provided.

[Ranking data storage unit 4]
The rank data storage 4 stores the rank data of the word (or noun) generated by the sentence analysis unit 2, and is provided as needed. In the data processing device 1 of the present embodiment, after the analysis processing by the sentence analysis unit 2, the short sentence creation unit 3 may continuously create a short sentence, but the result of analysis by the sentence analysis unit 2 is the ranking data storage unit 4. The short sentence creation unit 3 can also create a short sentence by using the data stored in the rank data storage unit 4.

[Event word database 5]
Nouns that can identify an event are recorded in the event word database 5, and are provided as necessary. Here, examples of nouns that can identify an event include nouns that are frequently used to represent events such as incidents, accidents, and disasters, nouns specific to the event, and nouns that represent regions. The nouns recorded in the event word database 5 are used when a noun representing a specific content such as a region or an event is selected in the word selection unit 31 or the like of the short sentence creation unit 3.

[motion]
Next, the operation of the data processing device 1 of the present embodiment, that is, a method of automatically creating a short sentence regarding a group of posted sentences by using the data processing device 1 will be described. In the data processing method of the present embodiment, a sentence analysis step and a short sentence creation step are performed by one or a plurality of data processing devices 1, and a short sentence related to a posted sentence group is automatically created. FIG. 2 is a flowchart showing the data processing method of the present embodiment, and FIG. 3 is an example of ranking data.

[Sentence analysis process]
The sentence analysis process is mainly performed by the article analysis unit 2 of the data processing device 1, analyzes each posted sentence of the posted sentence group related to an arbitrary event collected via the Internet, and the appearance frequency of words included in the posted sentence. Ranking by. The method of extracting and ranking words is not particularly limited, but for example, as shown in FIG. 2, after part-speech decomposition for each posted sentence (step S1), nouns are extracted from the part-speech-decomposed words. It can be carried out by (step S2) and ranking the extracted nouns according to the frequency of appearance (step S3).

Regarding the part-speech decomposition in step S1, the text data of each posted sentence in the posted sentence group can be divided into words by using an existing Japanese morphological analysis system or the like. As for the noun extraction in step S2, only the noun can be extracted from the part-speech-decomposed words by using machine learning or a Japanese morphological analysis system. Then, in the ranking of step S3, as shown in FIG. 3, the number of used words (appearance frequency) of the extracted nouns and the like may be counted, and the ranking may be made from the one with the largest number.

In the sentence analysis step, part of speech other than nouns such as adverbs may be extracted and ranked in combination with noun extraction or after noun extraction. Further, in the sentence analysis step, all the words decomposed by part of speech can be ranked, and only nouns can be extracted from the ranked words.

[Short sentence creation process]
The short sentence creation process is mainly performed by the short sentence creation unit 3 of the information processing apparatus 1, and creates a short sentence related to the posted sentence group based on the ranking data of nouns. The method is not particularly limited, but for example, as shown in FIG. 2, a noun representing a specific content is selected from the nouns included in each posted sentence of the posted sentence group (step S4). It can be carried out by selecting other part of speech such as an adverb that specifies a place or an event as necessary and creating a sentence using the selected noun or the like (step S5).

Specifically, in the noun selection step of step S4, nouns that can identify an event common to each posted sentence in the posted sentence group, such as "region" and "event", are given priority to those having a high rank. Select. The noun can be selected, for example, by comparing the data of the "noun that can specify the event" stored in the event word database 5 in advance with the ranking data of the noun.

As a result, for example, in the case of the ranking data shown in FIG. 3, "Fukuoka" is selected as the "region" and "accident" is selected as the "event". At this time, the highest-ranked "accident" is selected as the first candidate for the noun representing "event", but the next highest-ranked "earthquake" can also be selected as the second candidate.

Further, another part of speech that can identify the event may be selected in combination with the above-mentioned selection of the noun or after the selection of the noun. Specifically, an adverb (for example, "de") that is important when specifying a place or an event is selected by the same method as a noun. This makes it possible to create more natural short sentences.

Then, in the short sentence generation step of step S5, a short sentence related to the posted sentence group is created using the noun and other part of speech selected in the noun selection step. For example, in the case of the ranking data shown in FIG. 3, "there is a possibility of an accident in Fukuoka Prefecture", "there is a possibility of an earthquake in Fukuoka Prefecture", or "there is a possibility of an accident or an earthquake in Fukuoka Prefecture". A short sentence is created. As described above, in the data processing device 1 of the present embodiment, not only the noun having the highest rank but also a plurality of nouns having a high rank are selected, and the selected nouns are combined to create one short sentence. , You can also create multiple short sentences for each noun.

The sentence analysis step and the short sentence creation step described above may be performed continuously, but they may also be performed independently. For example, it is possible to carry out the sentence analysis process for all the posted sentences or the posted sentence group, and selectively carry out only the necessary short sentence creation steps. Further, in the above-mentioned short sentence creation step, a specific noun is selectively extracted from the words decomposed into part speech, but the present invention is not limited to this, and any word representing a posted sentence group is used. Words other than nouns may be used, and words other than nouns may be selected to create a short sentence.

Further, a short sentence (text data) related to the posted sentence group created by each of the above-mentioned steps is attached to the corresponding posted sentence group data (step S6) and distributed to the user, for example. Alternatively, apart from the posted text group data, only the created short text can be posted on the website or SNS and provided as information for a specific user.

[program]
Each of the above-mentioned steps can be carried out by creating a computer program for realizing each function of the data processing device and implementing it on one or more computers. That is, the data processing method of the present embodiment has a sentence analysis function that analyzes each posted sentence of a posted sentence group related to an arbitrary event collected via the Internet and ranks the words included in the posted sentence according to the frequency of appearance. It can be executed by a program that causes a computer to execute a short sentence creation unit that automatically creates short sentences related to a group of posted sentences based on word ranking data.

In this computer program, the sentence analysis function is realized by the function of part-speech decomposition for each posted sentence, the function of extracting nouns from the part-speech-decomposed words, and the function of ranking the extracted nouns by the frequency of appearance. You may. In addition, the short sentence creation function is selected by the word selection unit and the function to select a noun that represents a specific content from the nouns included in each posted sentence of the posted sentence group based on the ranking data of the word (or noun). It is realized by a sentence generation function that creates a short sentence using a noun, and in the function of selecting a noun, the noun with the higher rank may be preferentially selected.

It should be noted that each of the above-mentioned functions does not have to be installed in one program, and may be executed by creating a program for each function and linking them. In that case, each program may be divided into two or more computers or devices, mounted, and operated.

Since the posted text included in the posted text group contains various information other than the information related to a specific event, if you analyze the text data for each posted text and create a short sentence that identifies the event such as the title or heading. , Specific accuracy is reduced by information unrelated to the event. On the other hand, in the data processing device of the present embodiment, the words included in each posted sentence constituting the posted sentence group are ranked according to the frequency of appearance, and the sentence is created using the high-ranked words. It is possible to accurately identify the content and area of occurrence of the event that each posted text is talking about.

As a result, by using the data processing device, data, and data processing method of this embodiment, the headings and titles that accurately represent the contents of the posted texts related to incidents, accidents, disasters, etc. collected via the Internet. , Summary and introductory texts can be automatically created. In addition, when the data processing device of the present embodiment analyzes the latest posted texts on the SNS that have not been posted for a long time, it is necessary to quickly discover the events occurring in each place and their contents. It can be provided to people in a quick and easy-to-understand manner.

The data processing device of the present embodiment is particularly effective for processing a "posted sentence group" composed of two or more posted sentences related to an arbitrary event collected via the Internet, but of course, one post. It can also be applied to sentences. In that case, only one sentence is analyzed, words are ranked, and a short sentence related to the posted sentence is created based on the ranking data.

(Second embodiment)
Next, the data processing system according to the second embodiment of the present invention will be described. FIG. 4 is a conceptual diagram showing the configuration of the data processing system of the present embodiment. In the first embodiment described above, the analysis of each posted sentence of the posted sentence group and the creation of a short sentence are performed by the same device, but the present invention is not limited to this, and these steps are performed by different devices. It may be carried out. That is, as shown in FIG. 4, the data processing system 10 of the present embodiment includes a sentence analysis device 20 and a short sentence creation device 30.

[Sentence analysis device 20]
The sentence analysis device 20 analyzes each of the posted sentences 6a and 6b of the posted sentence group related to an arbitrary event collected via the Internet 5, and ranks the words included in the posted sentences 6a and 6b by the frequency of appearance. .. The configuration and operation of the text analysis device 20 other than the above are the same as those of the text analysis unit 2 of the first embodiment described above.

[Short sentence creation device 30]
The short sentence creation device 30 creates a short sentence related to the posted short sentence creation group based on the ranking data of the words created by the sentence analysis device 20. The configuration and operation of the sentence analysis device 30 other than the above are the same as those of the short sentence creation unit 3 of the first embodiment described above.

[Posted sentence classification device 7]
The data processing system 10 of the present embodiment may be provided with a posted text classification device 7 that classifies a plurality of posted texts collected via the Internet and creates a posted text group for each of the same events. In that case, the posted sentence group created by the posted sentence classification device 7 is input to the sentence analysis device 20, and the analysis process is performed.

In the data processing system 10 of the present embodiment, only the necessary ones among the posted sentence groups classified by the posted sentence classification device 7 may be output to the sentence analysis device 20 and analyzed. For example, it is also possible to monitor the occurrence of an event with the posted sentence classification device 7, collect data before and after the change at the timing when the change appears, classify the data, and send it to the sentence analysis device 20. As a result, data analysis can be performed efficiently.

Further, the data processing system 10 of the present embodiment may be provided with a distribution device that adds the text data created by the short text creation device 30 to the data of the posted text group and distributes it externally. In the data processing system 10 shown in FIG. 4, the posted sentence classification device 7 also serves as a distribution device, and the sentence created by the short sentence creating device 30 is added to the data of the posted sentence group by the posted sentence classification device 7. , Is delivered to the outside via the Internet 5.

The data processing system of the present embodiment collects posts related to the same event, ranks the words contained in the text data according to the frequency of appearance, and creates a sentence using the high-ranked words. It is possible to automatically create a short sentence that accurately represents it. The configurations and effects other than the above in this embodiment are the same as those in the first embodiment described above.

1 Data processing device 2 Sentence analysis unit 3 Short sentence creation unit 4 Rank data storage unit 5 Event word database 6a, 6b Contribution sentence 7 Contribution sentence classification device 8 Internet 10 Data processing system 20 Sentence analysis device 21 Part-of-word decomposition unit 22 Noun extraction unit 23 Noun count unit 30 Short sentence creation device 31 Word selection unit 32 Short sentence generation unit

Claims (13)

With a sentence analysis unit that analyzes each posted sentence of a posted sentence group consisting of two or more posted sentences related to the same event collected via the Internet, and ranks the words included in the posted sentence group by the frequency of appearance. ,
A short sentence creation unit that selects a word that can identify an event common to each posted sentence from the words included in the posted sentence group and automatically creates a short sentence that expresses the content of the event using the selected word.
An event word database in which nouns that can identify an event are stored and used when selecting a word that can identify an event in the short sentence creation unit.
Have,
The event is an incident, accident or disaster and
In the selection of words that can identify the event in the short sentence creation unit, a data processing device that preferentially selects words having a high frequency of appearance. The sentence analysis unit
The part-speech decomposition part that decomposes the part of speech for each post,
A noun extractor that extracts nouns from words that have been decomposed into parts of speech,
A noun count section that ranks the extracted nouns according to their frequency of appearance,
Equipped with
The data processing device according to claim 1, wherein the short sentence creating unit creates the short sentence using a noun having a high frequency of appearance. The short sentence making part
A word selection unit that selects nouns that can identify events common to each posted sentence in the posted sentence group from among the nouns included in the posted sentence group.
A short sentence generation unit that creates a short sentence using the nouns selected by the word selection unit, and
Equipped with
The word selection unit compares the noun data stored in the event word database with the noun ranking data created by the sentence analysis unit, and requests that the noun with the highest frequency of appearance be preferentially selected. Item 2. The data processing apparatus according to Item 1. The data processing device according to claim 3, wherein the noun that can identify the event is a noun representing a region and a noun representing an event. The data processing apparatus according to any one of claims 1 to 4, further comprising a ranking data storage unit for storing ranking data of the frequency of appearance of the word. The data processing apparatus according to any one of claims 1 to 5, wherein the short sentence is a title or a heading. A method of automatically creating a short sentence using one or more data processing devices.
With the data processing device
A sentence analysis process that analyzes each posted sentence of a posted sentence group consisting of two or more posted sentences related to the same event collected via the Internet, and ranks the words included in the posted sentence group by the frequency of appearance. ,
A short sentence creation process in which a word that can identify an event common to each posted sentence is selected from the words included in the posted sentence group and a short sentence expressing the content of the event is created using the selected word. Do,
The event is an incident, accident or disaster and
In the selection of a word that can specify the event in the short sentence creation step, a data processing method that preferentially selects a word having a high frequency of appearance by using an event word database in which a noun that can specify the event is stored . The sentence analysis step is
The process of part-speech decomposition for each posted sentence and
The process of extracting nouns from words that have been decomposed into parts of speech,
The process of ranking the extracted nouns by frequency of appearance,
Have,
The data processing method according to claim 7, wherein the short sentence creation step uses a noun having a high frequency of appearance to create the short sentence. The short sentence creation process is
The process of selecting a noun that can identify an event common to each posted sentence of the posted sentence group from the nouns included in the posted sentence group, and
The process of creating a short sentence using the selected nouns,
Have,
In the step of selecting the noun, the noun data accumulated in the event word database is compared with the noun rank data created in the sentence analysis step, and the noun with the highest frequency of appearance is preferentially selected. The data processing method according to claim 7 or 8. A sentence analysis device that analyzes each posted sentence of a posted sentence group consisting of two or more posted sentences related to the same event collected via the Internet, and ranks the words included in the posted sentence group by the frequency of appearance. ,
A short sentence creation device that selects a word that can identify the event common to each posted sentence from the words included in the posted sentence group, and automatically creates a short sentence that expresses the content of the event using the selected word. When,
An event word database in which nouns that can identify an event are stored and used when selecting a word that can identify an event in the short sentence creation device.
Have,
The event is an incident, accident or disaster and
A data processing system that preferentially selects words with a high frequency of appearance in the selection of words that can identify the event in the short sentence creation device. Equipped with a post text classification device that classifies multiple post texts collected via the Internet for each event and creates a post text group that is a collection of text data composed of two or more post texts related to the same event.
The data processing system according to claim 10, wherein the sentence analysis device analyzes a group of posted sentences created by the posted sentence classification device. The data processing system according to claim 10 or 11, further comprising a distribution device for externally distributing short sentence data created by the short sentence creation device by adding it to the data of the posted sentence group. On the computer
With a sentence analysis function that analyzes each posted sentence of a posted sentence group consisting of two or more posted sentences related to the same event collected via the Internet, and ranks the words included in the posted sentence group according to the frequency of appearance. ,
A short sentence creation function that selects a word that can identify the event common to each posted sentence from the words included in the posted sentence group, and automatically creates a short sentence that expresses the content of the event using the selected word. And execute,
The event is an incident, accident or disaster and
In the selection of words that can identify the event in the short sentence creation function, a program that preferentially selects words with a high frequency of appearance by using an event word database in which nouns that can identify the event are stored .

Images