JP4314568B2 - Television program search method and television program search device - Google Patents

Description

  The present invention learns a user's viewing tendency from past operation histories (program viewing, recording, playback, etc.) in an apparatus for receiving a television broadcast, and automatically extracts a user's favorite program based on the information. The present invention relates to a television program search method and a television program search apparatus for the above.

  With the widespread use of cable TV and digital broadcasting, a full-channel era has been put into practical use in Japan, and it is not easy for users to find a desired channel. In view of this, a program information guide called EPG (Electronic Program Guide) added to television broadcasting has been proposed to automatically record after extracting a favorite program based on a user's viewing history (for example, , See Patent Document 1.).

JP-A-5-62283

By the way, the method proposed in Patent Document 1 analyzes the user's preference from the viewing time of a program. However, there are many ways to see a general television program, and the program is casually watching the same program, and there are surprisingly few programs that are consciously trying to watch the program.
In the conventional method, the degree of preference increases because the frequency of viewing such programs is high, and conversely, the degree of preference decreases because the frequency of watching really interesting programs is low. There was a problem.

  The present invention has been made in view of the above points. A television program search method and a television program search apparatus capable of searching for a program corresponding to the user's original preference without being affected by the user's viewing frequency. The purpose is to provide.

In order to solve the above-mentioned problems, the present invention comprises means described in the following 1) to 2).
That is,
1) A television program search how to find the favorite program of Yu over The,
A program information acquisition step of acquiring and storing each program information corresponding to a program scheduled to be broadcast and a program already broadcast,
A first extraction step of extracting feature information indicating features of a program scheduled to be broadcast in a predetermined future time period from the stored program information ;
A second extraction step of extracting feature information indicating features of a program broadcast in the past predetermined time period from the stored program information ;
A first combination pattern generation step of obtaining all combination patterns in a permutation of programs scheduled to be broadcast in the predetermined future time period, and using each of the obtained combination patterns as a first combination pattern;
A second combination pattern generation step of generating a combination pattern of programs already broadcast in the past predetermined time period as a second combination pattern;
Based on the feature information of each program included in the first combination pattern and the feature information of each program included in the second combination pattern, all the first combination patterns and a plurality of the past predetermined information The degree of similarity with the second combination pattern in the time zone is obtained, and the second combination pattern in the first combination pattern and the second combination pattern that has obtained the largest degree of similarity among the obtained degrees of similarity. A selected program determining step of specifying a past predetermined time zone corresponding to a combination pattern and determining a program selected by the user in the specified past predetermined time zone by referring to an operation history related to the user's program selection; ,
A determination step of selecting a program scheduled to be broadcast in the future predetermined time zone having the same feature information as the feature information of the program selected by the user, and determining the selected program as a recommended program to the user;
A television program search method characterized by comprising:
2) A television program search device to search for favorite program of Yu over The,
Program information acquisition means for acquiring and storing each program information corresponding to a program scheduled to be broadcast and a program already broadcast,
First extraction means for extracting feature information indicating features of a program scheduled to be broadcast in a predetermined future time period from the stored program information ;
Second extraction means for extracting feature information indicating features of a program that has been broadcast in the past predetermined time period from the stored program information ;
First combination pattern generation means for obtaining all combination patterns in a permutation of programs scheduled to be broadcast in the predetermined future time period, and using each of the obtained combination patterns as a first combination pattern;
Second combination pattern generation means for generating a combination pattern of programs broadcast in the past predetermined time period as a second combination pattern;
Based on the feature information of each program included in the first combination pattern and the feature information of each program included in the second combination pattern, all the first combination patterns and a plurality of the past predetermined information The degree of similarity with the second combination pattern in the time zone is obtained, and the second combination pattern in the first combination pattern and the second combination pattern that has obtained the largest degree of similarity among the obtained degrees of similarity. A selected program determining means for specifying a predetermined past time zone corresponding to a combination pattern, and determining a program selected by the user in the specified past predetermined time zone by referring to an operation history related to the user's program selection; ,
Determining means for selecting a program scheduled to be broadcast in the future predetermined time zone having the same feature information as the feature information of the program selected by the user, and determining the selected program as a recommended program to the user;
A television program search device characterized by comprising:

According to the “television program search method and television program search device” of the present invention, it is possible to calculate a preference level that is not influenced by the user's viewing frequency, and to search for a program that is close to the user's original preference. A John program search method and a television program search apparatus can be provided.
In addition, since a program search is performed using a broadcast-specific characteristic that a program to be viewed is determined by selecting one of programs broadcast at a predetermined time, it is closer to the user's preference. It is possible to provide a television program search method and a television program search apparatus that can search for a program.
Further, since the user's viewing tendency is learned from the operation history of watching, recording, playing, etc. of the television program, the television program searching method and the television program that can search for a program closer to the user's preference A search device can be provided.

Hereinafter, the best mode for carrying out the invention of a television program search method and a television program search apparatus according to the present invention will be described with reference to preferred embodiments.
FIG. 1 is a schematic block diagram showing the configuration of a television program search apparatus applied to this embodiment. As shown in the figure, the television program search device 9 includes an EPG acquisition unit 1, a history processing unit 2, a storage device 3, extraction units 4 and 6, a generation unit 5, a determination unit 7, and a search unit 8. .

  The EPG information sent together with the television broadcast is supplied to the EPG processing unit 1 and sequentially stored in the storage device 3 in the device. Further, according to the operation from the operation supply unit 2, the history of each operation is sequentially stored in the storage device 3. The extraction unit 4 reads the program information of the program scheduled to be searched from the program information in the storage unit 3 and the program information of the program of another station broadcast at the same time from the storage device 3 and obtains a vector for each program. It supplies to the production | generation part 5. The generation unit 5 generates vectors corresponding to the number of combinations of programs from each supplied program and supplies the vectors to the search unit 8. At the same time, the extraction unit 6 reads the history from the operation history in the storage unit 3, vectorizes it together with the program information of the program broadcast at the same time as the history, and supplies it to the search unit 8. The search unit 8 calculates vectors and similarity for the number of combinations to be searched for each vector of the history, extracts a history having the largest similarity, and supplies the history to the determination unit 7. The determination unit 7 determines a recommended program from the program group to be searched based on the viewing channel for the supplied history.

  The program information actually handled in the television broadcast includes information such as a title name, a genre name, a performer name, and a program content commentary as shown in FIG. These pieces of information may be obtained from what is multiplexed with broadcast and sent as EPG information, or may be obtained from the Internet. In the future, it is also considered to be sent as more detailed data as metadata.

  In the present invention, the user's preference is determined based on the past operation history of the user and the degree of similarity between these pieces of information. Generally, the document unit (the present invention) is used when searching a document on the WEB. In this case, instead of making a determination based on the degree of similarity in units of programs), a program suited to the user's preference is extracted with higher accuracy by utilizing the unique characteristics of broadcasting.

  In general, when watching a television broadcast, an arbitrary program is not seen at an arbitrary time, and a program to be viewed is determined by always selecting one of the programs being broadcast at a certain time. . That is, for example, when a program as shown in FIG. 3 is broadcast at a certain time, only one of the three types of programs can be viewed. In addition, the selection method is often performed by relative comparison. Even if the variety shown in FIG. 3 is seen, the program shown in FIG. 8 is always watched at another time. It is common to watch baseball without watching variety.

Therefore, the present invention is characterized in that a process is performed for each combination of programs in order to determine a favorite program based on the similarity in units of program combinations rather than the similarity in units of programs.
Specific processing will be described below with reference to the flowchart of FIG.

  Program information of each program broadcast at a certain time is acquired from EPG information and the like and stored in the storage device 3. The extraction unit 4 reads program information of a program group broadcast in the time zone to be searched from the storage device 3 (STEP 1). Next, elements representing the characteristics of the program are extracted from the program information for each supplied program (STEP 2). For example, as shown in FIG. 2, elements such as a title name, a genre name, a performer name, and a commentary that are decomposed into morphemes can be considered.

  Next, each program is expressed as a vector with dimensions corresponding to the number of elements obtained by excluding duplicate elements from the elements of all supplied programs. In the case of FIG. 3, all elements that do not overlap are “news”, “variety”, and “drama”, so there are three elements and they are represented by a three-dimensional vector. The vector representation of each program is 100 for channel 1, 010 for channel 2, and 001 for channel 3 (STEP 3).

  Next, the first history is acquired from the operation history acquired in advance, and the program information of each program broadcast at the time corresponding to the history is acquired from the EPG information stored in advance, and the extraction unit 6 (STEP 4). Using the extracted elements, the extraction unit 6 vectorizes each program broadcast at the time corresponding to the history in the dimension obtained in STEP 3 (STEP 6). For example, in the example of the history shown in FIG. 4, it can be seen that the channel 1 was viewed from 19:00 to 20:00 on the day of month X. From FIG. 4, at the same time corresponding to the history, as shown in FIG. You can see that the program was broadcast. If each of these programs is expressed by a dimension vector generated in STEP 3, channel 1 is 100, channel 2 is 000, and channel 3 is 000.

  Next, the generation unit 5 obtains all combinations of the programs scheduled to be broadcast (channel 1, channel 2, channel 3). The combinations are (Channel 1, Channel 2, Channel 3), (Channel 1, Channel 3, Channel 2), (Channel 2, Channel 1, Channel 3), (Channel 2, Channel 3, Channel 1), (Channel 3, Channel 1, channel 2), (channel 3, channel 2, channel 1).

Next, the search unit 8 obtains the inner product of the program group vectors for the number of combinations with the program group vectors generated from the history (STEP 6 to STEP 9), and sets the largest one as the similarity to the history (STEP 10). . For example, each inner product for the above history is
(100010001) * (100000000) = 1
(100001010) · (100000000) = 1
(010100001) ・ (100000000) = 0
(010001100) · (100000000) = 0
(0011100010). (100000000) = 0
(001010100) · (100000000) = 0
Thus, the similarity is 1 when the sequence of programs is (channel 1, channel 2, channel 3) or (channel 1, channel 3, channel 2).

  Similarly, for the next history, when the similarity is obtained (STEP 11) and the arrangement order is (channel 2, channel 1, channel 3), the similarity is (010100001) · (010100001) = 3.

  In this way, the similarity is obtained for all the histories, and by finding the history having the maximum similarity, the history most similar to the broadcast-scheduled program group to be determined and its combination can be found (STEP 12). ~ 13).

In the determination unit 7, since the program viewed at that time from the history with the highest similarity obtained by the search unit 8 is “Channel 2”,
(Channel 1, Channel 2 , Channel 3) History list
↓ ↓ ↓
(Channel 2, Channel 1 , Channel 3) It is determined that the program most likely to be watched in the program group scheduled to be broadcasted is “Channel 1” based on the relationship of the search target (STEP 14).

  Further, considering that there is a program group as shown in FIG. 8 on another scheduled broadcast date, the vector elements of each program at that time are “news”, “variety”, and “baseball”. 100, channel 2 is 010, and channel 3 is 001. If these are processed in the same manner as in the previous example, the similarity to the third history in FIG. 4 is the highest, and the arrangement order at that time is (program 1, program 2, program 3). Since the program viewed from the third history is found to be “channel 3”, it is determined that the program most likely to be viewed in the broadcast scheduled program group is “channel 3”.

  Thus, according to the present invention, there is provided a television program search method and a television program search device that can determine a user's preference more accurately regardless of viewing frequency. Also, a television broadcast receiving apparatus having a function for guiding a user to quickly select a program searched by the television program searching method and the television program searching apparatus according to the present invention, or a searched program It is also possible to construct an automatic recording device that always records a user's favorite program by automatically recording

It is the schematic block diagram which showed the structure of the television program search apparatus applied to a present Example. It is the figure which showed the example of the program information used for television broadcasting. It is the figure which showed the example of the program currently broadcast at the predetermined time. It is the figure which showed an example of the user's viewing history. It is the figure which showed the example of the program currently broadcast at the predetermined time. It is the figure which showed the example of the program currently broadcast at the predetermined time. It is the figure which showed the example of the program currently broadcast at the predetermined time. It is the figure which showed the example of the program which is scheduled to be broadcast at predetermined time. It is the figure which showed an example of the flowchart which showed the television program search method applied to a present Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... EPG acquisition part 2 ... History processing part 3 ... Accumulation apparatus 4,6 ... Extraction part 5 ... Generation part 7 ... Determination part 8 ... Search part 9 ... Television program search apparatus

Claims (2)

A television program search how to find the favorite program of Yu over The,
A program information acquisition step of acquiring and storing each program information corresponding to a program scheduled to be broadcast and a program already broadcast,
A first extraction step of extracting feature information indicating features of a program scheduled to be broadcast in a predetermined future time period from the stored program information ;
A second extraction step of extracting feature information indicating features of a program broadcast in the past predetermined time period from the stored program information ;
A first combination pattern generation step of obtaining all combination patterns in a permutation of programs scheduled to be broadcast in the predetermined future time period, and using each of the obtained combination patterns as a first combination pattern;
A second combination pattern generation step of generating a combination pattern of programs already broadcast in the past predetermined time period as a second combination pattern;
Based on the feature information of each program included in the first combination pattern and the feature information of each program included in the second combination pattern, all the first combination patterns and a plurality of the past predetermined information The degree of similarity with the second combination pattern in the time zone is obtained, and the second combination pattern in the first combination pattern and the second combination pattern that has obtained the largest degree of similarity among the obtained degrees of similarity. A selected program determining step of specifying a past predetermined time zone corresponding to a combination pattern and determining a program selected by the user in the specified past predetermined time zone by referring to an operation history related to the user's program selection; ,
A determination step of selecting a program scheduled to be broadcast in the future predetermined time zone having the same feature information as the feature information of the program selected by the user, and determining the selected program as a recommended program to the user;
A television program search method characterized by comprising:
A television program search device to search for favorite program of Yu over The,
Program information acquisition means for acquiring and storing each program information corresponding to a program scheduled to be broadcast and a program already broadcast,
First extraction means for extracting feature information indicating features of a program scheduled to be broadcast in a predetermined future time period from the stored program information ;
Second extraction means for extracting feature information indicating features of a program that has been broadcast in the past predetermined time period from the stored program information ;
First combination pattern generation means for obtaining all combination patterns in a permutation of programs scheduled to be broadcast in the predetermined future time period, and using each of the obtained combination patterns as a first combination pattern;
Second combination pattern generation means for generating a combination pattern of programs broadcast in the past predetermined time period as a second combination pattern;
Based on the feature information of each program included in the first combination pattern and the feature information of each program included in the second combination pattern, all the first combination patterns and a plurality of the past predetermined information The degree of similarity with the second combination pattern in the time zone is obtained, and the second combination pattern in the first combination pattern and the second combination pattern that has obtained the largest degree of similarity among the obtained degrees of similarity. A selected program determining means for specifying a predetermined past time zone corresponding to a combination pattern, and determining a program selected by the user in the specified past predetermined time zone by referring to an operation history related to the user's program selection; ,
Determining means for selecting a program scheduled to be broadcast in the future predetermined time zone having the same feature information as the feature information of the program selected by the user, and determining the selected program as a recommended program to the user;
A television program search device characterized by comprising:

Images