JP2001101195A - Device for retrieving television program and device for automatically selecting program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To effectively retrieve a television program related to news topics, and to present the program and general information such as a popular program and a program having popularity in a television program retrieving device which utilizes a news sentence and an explanatory note of a television program and an automatic program selecting device which automatically selects and presents a television program. SOLUTION: These devices are provided with topic representing means 11 and 12 which utilize change with time of the frequency in use of words in news sentences and generate data representing news topics.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a television program search device using a news text and a description of a television program, and an automatic program selection device for automatically selecting and presenting a television program.

[Summary of the Invention] [0002] The present invention utilizes a data structure accurately expressing a news topic as a search condition in a search for a television program, and a data structure expressing a description of a television program as a search target. By using it, a TV program related to a news topic can be effectively searched. In addition, using the search results, the automatic program selection function that automatically selects and presents a large number of programs for individual viewers of the TV has become a topic of public attention. The program information that specializes in "" is added to expand the range of selection.

[0003]

2. Description of the Related Art In general, in information retrieval for sentences,
An important word serving as a keyword is extracted from a sentence, and a search is performed by comparing the word with a word serving as a search condition. In such a method, it is difficult to select a word to be given as a search condition, and a sentence to be searched cannot be accurately expressed.
As a result, unnecessary sentences may be output, or necessary sentences may not be output. In an automatic program selection method for selecting and providing a favorite program to individual viewers, information describing the contents of a program (hereinafter, referred to as a “program index”) and an individual describing the viewer's preference are described. The program is selected by comparing keywords included in the information (hereinafter, referred to as “user profile”).

[0004]

However, when searching for a TV program from a news topic, a word is used as a search condition in the conventional search method, but the content of the news topic cannot be accurately expressed. In addition, good results cannot be obtained.

[0005] Conventionally, a search is performed by matching a word serving as a search condition with a word included in an explanatory sentence to be searched. However, in this method, since there are many words (words with a high frequency but a small amount of information) that have an adverse effect in natural language processing, it is difficult to accurately evaluate the similarity between the search condition and the search target.

Further, when automatically selecting and presenting a favorite program to individual viewers, a user profile and a program index are used. However, in this method, program selection is biased and the program is popular. General information such as programs and topical programs cannot be used.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has a general purpose such as a television program search apparatus capable of effectively searching for a television program related to a news topic and a popular program or a topical program. It is an object of the present invention to provide an automatic program selection device capable of presenting additional information.

[0008]

According to a first aspect of the present invention, there is provided a television program search apparatus for a news topic utilizing a time change of a frequency of appearance of a word in a news sentence. And a topic expression means for generating data expressing the expression.

That is, according to the first aspect of the present invention, the data to be generated is generated by using the temporal change of the appearance frequency. Therefore, the topic of the news, which is a search condition, can be accurately expressed.

According to a second aspect of the present invention, there is provided a television program search device comprising a television program expression means for generating data representing a television program by using word appearance information included in a description of the television program. It is characterized by:

That is, according to the second aspect of the present invention, since the generated data is generated using the appearance information, the television program to be searched can be accurately represented.

According to a third aspect of the present invention, there is provided a television program search apparatus, comprising: topic expression means for generating data expressing a topic of a news using a time change of the frequency of appearance of a word in a news sentence; TV program expression means for generating data representing a TV program using the appearance information of words included in the description, and similarity between the data representing the topic and the data representing the TV program, Search means for searching for TV programs related to the topic;
It is characterized by having.

That is, according to the third aspect of the present invention, when the similarity between the topic of the news and the data of the description of the television program is equal to or more than a certain value, the similarity is determined to be related, and the data is searched. Therefore, an accurate search can be performed.

According to a fourth aspect of the present invention, there is provided the television program search device according to the first or third aspect, wherein the topic expression means is provided for each time of the appearance frequency of a word in a news sentence. Topic extracting means for extracting a topic of a news based on the degree of importance changing to a news group, and a news group forming the extracted topic is determined by a word weighted by the importance included in the news group and a contribution degree of the word. And a feature extracting means for expressing the generated data.

That is, in the invention according to claim 4, the topic expression means expresses the news group forming the extracted topic by data using words weighted by importance that changes with time. Therefore, the topic of the news, which is a search condition, can be accurately expressed.

According to a fifth aspect of the present invention, there is provided an automatic program selecting apparatus, wherein a topic expressing means for generating data expressing a topic of a news by using a time change of a frequency of appearance of a word in a news sentence; TV program expression means for generating data representing a TV program using the appearance information of words included in the description, and similarity between the data representing the topic and the data representing the TV program, Search means for searching for TV programs related to the topic;
When automatically selecting a program from a large number of television programs and presenting the program to individual viewers, a presentation means for additionally presenting the searched television program information is provided.

That is, according to the fifth aspect of the present invention, when a favorite program is selected and presented to individual viewers, a program that takes up an event that has become a topic in the world is also recommended to the viewers. And the range of program selection can be expanded.

An automatic program selection device according to a sixth aspect of the present invention is the automatic program selection device according to the fifth aspect,
And an ordering means for ordering the searched television programs by topicality and giving the order to the presentation means.

That is, according to the invention described in claim 6, since a program retrieved from an upper topic can be determined to be a program that has taken up a topic of the world, an accurate program selection can be performed.

According to a seventh aspect of the present invention, in the automatic program selection apparatus according to the fifth or sixth aspect, the topic expression means is provided for each time of the appearance frequency of a word in a news sentence. Topic extracting means for extracting a topic of a news based on the degree of importance changing to a news group, and a news group forming the extracted topic is determined by a word weighted by the importance included in the news group and a contribution degree of the word. And a feature extracting means for expressing the generated data.

[0021]

FIG. 1 is a block diagram showing the configuration of an embodiment of a television program search device and an automatic program selection device according to the present invention.

In FIG. 1, a television program search device 1 of the present embodiment receives a news topic extraction unit 11 for inputting a news manuscript as a search condition and an output (topic within a specific period) of the news topic extraction unit 11. Topic feature extraction unit 12
A TV program feature extraction unit 13 to which a description of a TV program to be searched is input; an output of a topic feature extraction unit 12 (topic feature vector); and an output of the TV program feature extraction unit 13 (a TV program description And a TV program search unit 14 that receives a search result (a TV program related to a topic) upon receipt of each of the feature vectors.

Further, the automatic program selection device 2 of the present embodiment
Is a program editing unit 2 receiving the output (search result) of the TV program search unit 14 in addition to the configuration of the TV program search device 1 described above.
5 and a program selection unit 26 that automatically selects and presents a program to a viewer in response to an output of the program editing unit 25 (a television program related to an ordered topic).

The correspondence between the above configuration and the claims is as follows. The topic expression means includes a news topic extraction unit 1
1 and the entire topic feature extraction unit 12 correspond to each other. The TV program feature extraction unit 13 corresponds to the TV program expression means.
The TV program search section 14 corresponds to the search means. The program selection unit 26 corresponds to the presentation means. The program editing unit 25 corresponds to the ordering means.

The operation of the television program search device and the automatic program selection device of the present embodiment will be described below with reference to FIGS. FIG. 2 is an example of topic extraction. 3 and 4 are examples of news manuscripts forming one topic.
FIG. 5 is an example of a topic feature vector. FIG. 6 is an example of a description of a television program. FIG. 7 is an example of a result of searching for a TV program from news topics.

First, the news topic extracting unit 11 obtains, for example, a monthly importance Weight (t) of a word included in the news manuscript from the input electronic manuscript manuscript by the following equation (1). Based on this, the articles included in the news manuscript are classified into similar items, and topics are extracted.

[0027]

(Equation 1) In the equation (1), n is the frequency at which the word t appears in a certain month, and e is the expected value. N is the total number of news manuscripts in the target month, and df (t) is the number of manuscripts in which the word t appears.

FIG. 2 shows an example of topic extraction. One topic is formed by a plurality of news manuscripts. For example, the topic "missile launches of North Korea and the Democratic People's Republic of Korea" in FIG. 2 is formed by a large number of news manuscripts as shown in FIGS. That is, in the topic extraction example shown in FIG. 2, more accurate topic extraction is actually performed.

To search for a television program from this news topic, data (feature vector) characterizing the topic is used.
And data (feature vector) characterizing the description of the television program. Therefore, the topic feature extraction unit 12 expresses a news group forming one topic by a feature vector using words weighted by importance that changes with time.

More specifically, the topic feature extraction unit 12 extracts a word in a news manuscript included in a topic extracted within a specific period by the news topic extraction unit 11, and uses the word as a feature vector element. Also, the contribution of the word to the topic
The appearance rate of the word in the news manuscript group in the topic corresponding to the word and the importance determined by equation (1) are defined by equation (2) as an operation element, and the value of the contribution is defined as the value of the element of the feature vector. .

[0031]

(Equation 2)

FIG. 5 is an example of a feature vector of the topic "missile launches of North Korea and the Democratic People's Republic of Korea".
In FIG. 5, the elements of the vector are words that have appeared in the news group constituting this topic. The value of the element of the vector is the value indicated by the number in parentheses in FIG. 5, which is the value of equation (2).

The TV program feature extraction unit 13 expresses a description of a TV program to be input as a search target by a feature vector using appearance information of words contained therein. Specifically, FIG. 6 shows an example of a description of a television program. The television program feature extraction unit 13 extracts a word from this description and uses the word as an element of a feature vector. Well, the values of the elements of the vector are not defined because they are not used in the processing here.

Next, the TV program search unit 14 extracts the news topic feature vector extracted by the topic feature extraction unit 12 and the feature vector of the TV program description sentence extracted by the TV program feature extraction unit 13. Use it to search for TV program descriptions related to news topics.

More specifically, the similarity between the news topic and the description of the television program is defined by equation (3).

[0036]

(Equation 3)

The TV program search section 14 calculates the similarity from all news topics to the explanations of all TV programs from the equation (3), and when the similarity is equal to or more than a certain value, for example, the similarity When> 0.5, it is determined that the topic of the news and the description of the television program are deeply related, and that is set as a search result. FIG. 7 shows an example of a result of searching for a television program from a topic of news. The above is the operation of the television program search device 1.

Next, the program editing section 25 orders the programs related to the topic. At this time, the TV programs related to the topic searched as described above are ranked according to the topicality. . Here, it is determined that a program retrieved from a topic having a large sum of element values of the feature vector is a program that has taken up a topic of the world.

As a result, when automatically selecting and presenting a television program to each viewer, the program selecting section 26 can additionally present a television program related to the ranked topic.

As a result, in the program selection section 26, when an automatic program is selected for an existing individual viewer, the program has become a hot topic besides the program selected by matching the keywords described in the user profile. Programs highlighting the event can also be recommended to viewers, expanding the range of program choices.

[0041]

As described above, according to the first to fourth aspects of the present invention, it is possible to generate data in which a news topic is accurately expressed. The TV program to be searched can be accurately searched from the similarity with the data that accurately expresses. As a result, it is possible to automate the association of news with a television program. For example, it is possible to automatically generate an advertisement for a television program related to the news during the news.

Further, according to the fifth and sixth aspects of the present invention, when automatically selecting and presenting a favorite program of each viewer, a program featuring an event that has become a hot topic in the world Since additional information can be presented, the range of program selection can be expanded.

[Brief description of the drawings]

FIG. 1 is a configuration block diagram showing an embodiment of a television program search device and an automatic program selection device according to the present invention.

FIG. 2 is an explanatory diagram illustrating an example of topic extraction.

FIG. 3 is an explanatory diagram showing an example of a news manuscript forming one topic (part 1).

FIG. 4 is an explanatory view showing an example of a news manuscript forming one topic (part 2).

FIG. 5 is an explanatory diagram showing an example of a topic feature vector.

FIG. 6 is an explanatory diagram showing an example of a description sentence of a television program.

FIG. 7 is an explanatory diagram showing an example of a result of searching for a TV program from a topic of news.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 TV program search device 1 2 Automatic program selection device 11 News topic extraction unit 12 Topic feature extraction unit 13 TV program feature extraction unit 14 TV program search unit 25 Program editing unit 26 Program selection unit

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Masahiro Shibata F-term in the Japan Broadcasting Corporation, Research Institute of Broadcasting Technology 1-10-11 Kinuta, Setagaya-ku, Tokyo 5B075 ND03 ND20 NK10 PR08 QM10 5C025 AA23 BA11 BA18 BA25 DA10

Claims (7)

[Claims] 1. A television program search device, comprising: topic expression means for generating data expressing a topic of a news using a temporal change of a frequency of appearance of a word in a news sentence. 2. A television program search device, comprising: television program expression means for generating data representing a television program by using appearance information of words included in a description of the television program. 3. A topic expression means for generating data expressing a topic of a news using a temporal change in the frequency of appearance of a word in a news sentence, and using appearance information of a word included in a description of a television program. TV program expression means for generating data representing a TV program, and search means for searching for a TV program related to the topic from a similarity between the data representing the topic and the data representing the TV program, A television program search device comprising: 4. The television program search device according to claim 1, wherein the topic expression means extracts a topic of the news based on a degree of importance of the appearance frequency of a word in a news sentence that changes with each period. Topic extracting means; and feature extracting means for expressing a news group forming the extracted topic by a word included in the news group and weighted by the importance and data generated by the degree of contribution of the word. A television program search device characterized by the above-mentioned. 5. A topic expression means for generating data expressing a topic of a news using a temporal change in the frequency of appearance of a word in a news sentence, and using appearance information of a word included in a description of a television program. TV program expression means for generating data representing a TV program, and search means for searching for a TV program related to the topic from a similarity between the data representing the topic and the data representing the TV program, Presenting means for adding and presenting the searched television program information when automatically selecting a program from a large number of television programs and presenting the program to individual viewers. Selection device. 6. The automatic program selection device according to claim 5, further comprising: an ordering unit that orders the searched TV programs by topicality and gives the order to the presentation unit. 7. The automatic program selection device according to claim 5, wherein the topic expression means extracts a topic of the news on the basis of a degree of importance of the appearance frequency of a word in a news sentence that changes at each time. Topic extracting means; and feature extracting means for expressing a news group forming the extracted topic by a word included in the news group and weighted by the importance and data generated by the degree of contribution of the word. An automatic program selection device, characterized in that: