JP4730869B2 - Television program information display device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention is an apparatus for displaying program information of a television program using electronic program guide (EPG data) data, and has a function of searching for and presenting a program. Relates to the device.
[0002]
[Prior art]
In television broadcasting or the like, the broadcast schedule for each program is determined in advance. In recent years, the broadcast schedule is distributed in various electronic media in the form of EPG data. The EPG data is distributed using, for example, the vertical blanking period together with the broadcast wave of a television program, and therefore the distributed EPG data can be acquired by a broadcast receiving means for receiving each program. it can. Further, EPG data is distributed regularly or irregularly from a service provider via a network such as the Internet, and can be obtained by communication means such as a telephone line. The obtained EPG data is Displayed on the display device of the personal computer.
[0003]
The program information of the television program based on the EPG data is displayed by a program information display device such as a television device. The program information display device displays the title name of the television program, the explanatory text of the program content, the broadcast date and time, and the broadcast station name based on the EPG data. Various program search means are used to search for a desired program. For example, a specific program can be searched based on additional information such as a genre type of a program and a performer name. Such program search means has been improved in order to improve convenience for the user.
[0004]
For example, in Japanese Patent Laid-Open No. 2000-67074, when a user reserves a program, the program content explanation text of the program included in the EPG data is displayed, and a word contained in the program content explanation text is designated as a keyword. Thus, a configuration for searching for a program related to a designated program is disclosed. Thus, when a program is reserved, the user can search for a program related to the reserved program by specifying a word in the program description text of the reserved program. There is no need to set a keyword for specifying a program, and operability is improved.
[0005]
Furthermore, Japanese Patent Laid-Open No. 11-8808 discloses a configuration for searching for a specific program by defining attribute information characterizing the program, classifying the program in advance based on the attribute information, and specifying a condition regarding the attribute information. Is disclosed. With such a configuration, a specific program can be searched at high speed.
[0006]
[Problems to be solved by the invention]
However, in these program search means, the specification of conditions may be further complicated in order to search for a specific program more accurately.
[0007]
For example, in the configuration disclosed in Japanese Patent Laid-Open No. 2000-67074, it is necessary for the user to select a keyword suitable for the search from the program content explanation text included in the EPG data. Therefore, the word selected by the user is not necessarily appropriate as a keyword, and the target program may not be searched.
[0008]
Further, in the configuration disclosed in Japanese Patent Laid-Open No. 11-8808, the user can specify the relationship between predetermined attribute information and a program, that is, what conditions are specified for the various attribute information. For example, it is necessary to understand whether a desired program can be searched. If such an understanding is not sufficient, there is a possibility that the user may not know how to specify the condition related to the attribute information for the program to be searched. In addition, there may be cases where attribute information of a program that is considered appropriate as a program searched by the user is not set. For example, even if the user wants to search for “bright atmosphere program”, “bright atmosphere” may not be set as attribute information. In this case, the user specifies the search condition. Is not easy.
[0009]
The user of the apparatus for displaying the program information of the television program is not a special person having specialized knowledge but a general person who views the television program in a general home. For this reason, as exemplified here, the method for specifying the conditions for searching for a program becomes complicated, and as a result, the user may not be able to use the program easily.
[0010]
The present invention solves such problems of the prior art, and provides a program information display device that allows ordinary people who do not have special knowledge to easily search for a specific program from EPG data. For the purpose.
[0011]
[Means for Solving the Problems]
The program information display device of the present invention includes a program guide acquisition means for acquiring electronic program guide data distributed together with a broadcast wave of a television program or distributed over a network, and the acquired electronic program guide data in a database. As storage means, program information display means for displaying program information included in the database, program designation means for designating an arbitrary program by a user from program information displayed on the program information display means, , Feature extraction means for extracting program features from program information included in the database, similarity calculation means for comparing features of a plurality of programs to obtain similarity between programs, and a plurality of features obtained by the feature extraction means Based on the characteristics of the program, the similarity calculation means is used to compare the similarities between the programs, and similar programs related to the program designated by the program designation means are obtained. Searching program search means, search result display means for displaying search results, program information display means, program specifying means, program search means and search result display means are controlled by the program information display means. Related program display control for displaying program information, causing the program search means to search for a program having a high similarity to the program specified by the user by the program specifying means, and displaying the search result on the search result display means With means The feature of the program extracted by the feature extraction means is the description of the program content obtained by calculating the sum of the semantic vector values respectively corresponding to the words included in the description text of the program content included in the program information. It is a semantic vector value obtained by normalizing the semantic vector value of a sentence, and the semantic vector value is a multi-dimensional vector value in which one dimension is assigned to each concept. A description content storing means for storing a description content of the program content for explaining the program content included in the program and a program ID of the corresponding program, a word constituting each description content of the program content, and the word A word storage means for storing a list of program IDs of programs corresponding to the explanatory text of the program content including the word, and a word semantic case for storing a semantic vector value representing the meaning of the word together with the word Yes and means As a result, the above object is achieved.
[0012]
The storage means Number Word extraction means for decomposing the explanatory text of the set contents into words; The above Explanation text of all program contents stored in the explanation storage means Distribution of words in Based on the meaning of individual words By representing the meaning vector value, the meaning of the word Quantification I do Meaning quantification means And It can be set as the structure which has.
[0013]
The feature extracting means includes a word extracting means for decomposing an explanatory sentence of program contents included in program information into words, and a meaning of each decomposed word The meaning vector value obtained by quantifying the meaning of the word in the word meaning storage means of the storage means The database Read from Word meaning acquisition means; Obtained by quantification in the semantic quantification means of the storage means, Meaning of individual words By operation on the semantic vector value corresponding to , Which means the meaning of the explanatory text of the program content consisting of individual words Find semantic vector values A sentence meaning obtaining means and a sentence feature extracting means for extracting the numerical meaning of the meaning of the explanation text of the program contents as the feature of the sentence can be adopted.
[0014]
The program search means may include a similarity acquisition means for obtaining a similarity between programs based on characteristics of a plurality of programs, and a similarity order arranging means for arranging a plurality of programs in descending order of similarity. it can.
[0015]
The storage means has genre information storage means for storing genre information of programs, and the feature extraction means is genre information of programs. By assigning the semantic vector to the genre information of the program It can be set as the structure which has the genre information addition means which digitizes and adds to the feature of a program.
[0016]
The genre information adding means includes a genre-specific fixed weight storage means for storing a fixed weight value for each genre, and a genre information for each genre for obtaining a weight based on genre information for each program using the fixed weight value. It can be set as the structure which has a weight addition means.
[0017]
The genre information adding means is the database. Using the ratio of the number of programs belonging to the genre to the total number of programs included in A genre weight calculating unit that calculates a dynamic weight value for each genre and a genre weight adding unit that obtains a weight based on genre information for each program using the calculated weight value. be able to.
[0018]
The genre information adding means includes weight combining means for combining a fixed weight value for each genre and a dynamic weight value calculated using a database, and weighting is performed using the combined weight value. The performed genre information may be added to the program characteristics.
[0019]
The program designating unit weights program information of a plurality of designated programs and a plurality of program designating units by which a user designates a plurality of arbitrary programs from program information displayed on the program information display unit. With synthesis means A program weight specifying means for the user to specify the weight of a plurality of arbitrary programs; Have The feature synthesizing unit attaches a designated weight to program information of a plurality of designated programs. It can be configured.
[0021]
The operation of the present invention will be described below.
[0022]
In the present invention, the program information included in the acquired EPG data is displayed on the program information display means, and the user designates an arbitrary program from the displayed information by the program designation means. It is possible to search for similar programs highly relevant to the designated program using the program characteristics as search conditions. Furthermore, since the program features are extracted by using the description text of the program content and the genre information included in the acquired EPG data, it is possible to obtain appropriate program features as search conditions. Furthermore, it is possible to modify the characteristics of the program by using the magnitude of the influence of the genre information on the program as a weight, and to correct the search condition to a more appropriate one.
[0023]
In the present invention, it is possible for a user to search for highly relevant programs having characteristics common to a plurality of designated programs by designating a plurality of arbitrary programs. Furthermore, the user can specify weight values for a plurality of programs, modify the characteristics of the programs with the specified weight values, and modify the search conditions to be more appropriate.
[0024]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
[0025]
FIG. 1 is a block diagram showing a configuration of a program information display apparatus according to an embodiment of the present invention. In addition, the arrow of FIG. 1 represents the data flow between each means which comprises the program information display apparatus of this embodiment. Also, the two broken arrows a and b in FIG. 1 indicate the relationship between the flow of processing in the present embodiment, which will be described later with reference to FIGS. 7 and 8, and each means constituting the program information display device of the present embodiment. Respectively.
[0026]
This program information display device has a program guide acquisition means 1 that is means for acquiring electronic program guide (EPG data) data to be distributed, and a related program display control means 9 that controls the program guide acquisition means 1 and the like. ing.
[0027]
The EPG data includes, for example, program information such as broadcast start date / time, broadcast end date / time, broadcast station name, program title name, program content explanation text, etc., and also attribute information such as genre information. The EPG data is distributed from a broadcasting station, a service provider having television program information, or the like. As a delivery method of EPG data, for example, there are a method of delivering using a vertical blanking period together with a broadcast wave of a television program, a method of delivering regularly or irregularly via a network, and the like. The program guide acquisition unit 1 is configured by, for example, a tuner that receives broadcast waves, a CPU connected to a network, and the like.
[0028]
The related program display control means 9 controls each means such as a program guide acquisition means when performing processing to search for and display related programs, and the EPG data acquired by the program acquisition means 1 is related program display control means. 9 is analyzed and output to the storage means (database) 2.
[0029]
The storage means 2 stores the EPG data acquired by the program guide acquisition means 1 in a database. As the storage means 2, for example, a semiconductor memory, a magnetic disk, or the like is used. A detailed configuration of the storage unit 2 will be described later with reference to FIG. Program information stored as a database in the storage unit 2 is output to the program information display unit 3 in accordance with an instruction from the related program display control unit 9.
[0030]
The related program display control means 9 reads out program information of a plurality of programs stored in the storage means 2 as a database and displays them on the program information display means 3. The program information display means 3 is constituted by, for example, a CRT, a liquid crystal display device or the like. When program information related to a plurality of programs is displayed by the program information display means 3, program information of a specific program is designated by the program designation means 4.
[0031]
The program designating unit 4 is configured by, for example, a mechanical operation unit such as a button or a switch, or a GUI (Graphical User Interface) displayed together with program information on the program information display unit 3. Used when selecting a program. A detailed configuration of the program designating unit 4 will be described later with reference to FIG. When program information of a specific program is specified by the program specifying unit 4, the specified program information is output to the feature extracting unit 5 by the related program display control unit 9.
[0032]
The feature extraction unit 5 obtains the feature of the program based on the program information of the program designated by the program designation unit 4. A more detailed configuration of the feature extraction unit 5 will be described later with reference to FIG. The feature of the program extracted by the feature extraction unit 5 is output to the program search unit 7.
[0033]
The program search means 7 is provided with a similarity calculation means 6 for searching for similar programs. The similarity calculation means 6 compares the features of a plurality of programs obtained by the feature extraction means 5 and compares them. Find out how much similarity between programs. The similarity between programs calculated by the similarity calculation means 6 is output to the program search means 7.
[0034]
The program search means 7 compares similarities between programs using the similarity calculation means 6 based on the characteristics of the plurality of programs obtained by the feature extraction means 5 and searches for programs having similar characteristics. A more detailed configuration of the program search means 6 will be described later with reference to FIG. The search result obtained by the program search means 7 is output to the search result display means 8.
[0035]
On the search result display means 8, program information of similar programs is displayed by the related program display control means 9 as a search result obtained by the program search means 7. The search result display means 8 is composed of, for example, a CRT, a liquid crystal display device or the like.
[0036]
FIG. 2 is a block diagram for explaining the database stored in the storage unit 2 in this embodiment in more detail.
[0037]
The storage means 2 includes explanation text storage means 10 for storing all program content explanation texts included in the program information acquired by the program guide acquisition means 1, and genre information for storing genre information of all programs included in the program information. Storage means 15. The program explanation text is a text explaining the program contents of each program. As for the genre information, the genre to which each program belongs is set for each program, and the genre information storage means 15 stores the genre information of each program for each program.
[0038]
The explanatory text storage means 10 is constituted by, for example, a semiconductor memory, a magnetic disk, or the like. The program content explanatory text stored in the explanatory text storage means 10 is output to the program information display means 3, and the program content explanatory text stored in the explanatory text storage means 10 is used when extracting program characteristics. It is output to the feature extraction means 5.
[0039]
FIG. 12 shows an example of a method for storing the program content explanation text by the explanation text storage means 10. When storing the program content explanatory text in the explanatory text storage means 10, unique ID numbers (0000001, 0000002,...) Are assigned to all programs included in the program information acquired by the program guide acquisition means 1. Are stored, and the ID number and the program content explanation text for the program with the ID number are stored as a set. The program content explanation text is managed by an ID number assigned to each program.
[0040]
In this example, as the minimum information in one program, the ID number of the program and the program content explanation text are stored as one set for one program. However, in consideration of user convenience For one program, the broadcast station name, broadcast date and time, etc. of the program may be stored as a set together with the program ID and the program content explanation text.
[0041]
As the genre information storage means 15, for example, a semiconductor memory, a magnetic disk or the like is used. The type of genre to which the program stored in the genre information storage unit 15 belongs is used by the feature extraction unit 5 when extracting the feature of the program.
[0042]
FIG. 16 is an explanatory diagram showing an example of a method for storing the genre type for each program in the genre information storage means 15. Even when the genre type for each program is stored in the genre information storage unit 15, unique ID numbers (0000001, 0000002, and so on) assigned to all programs included in the program information acquired by the program guide acquisition unit 1 ..) And genre information for a program to which an ID number is assigned are stored as one set. Genre information is managed by an ID number assigned to each program.
[0043]
In this example, in the distributed EPG data, it is assumed that one program belongs to one genre type, and only one genre type is given to one program. However, for example, even when two or more genre information is given to one program, the genre information can be handled in the same manner.
[0044]
The storage means 2 is provided with word extraction means 11. The word extracting means 11 is configured to cut the program content explanatory text stored in the explanatory text storage means 10 into words and store them in the word storage means 12.
[0045]
The word extracted by the word extraction unit 11 is stored in the word storage unit 12 in association with the program content explanation text that includes the word. The word storage unit 12 is configured by, for example, a semiconductor memory, a magnetic disk, or the like.
[0046]
For example, a morphological analysis technique can be used as a method of dividing the program content explanation sentence for each word. The morphological analysis technique is a technique widely used when performing Japanese processing such as Kana-Kanji conversion processing, and analyzes Japanese sentences and separates them into parts of speech. In this embodiment, nouns, verbs, and adjectives are set as parts of speech that reflect the semantic content of sentences. Note that the types of parts of speech to be extracted are not necessarily limited to these. For example, other parts of speech such as adjective verbs may be extracted.
[0047]
FIG. 13 is an explanatory diagram of an example of a method for storing words by the word storage unit 12. Here, for one word included in all program content explanation texts, ID numbers of all programs including the word are collectively stored as a program ID list.
[0048]
The storage means 2 is further provided with a semantic digitizing means 13. Based on the program content explanation text stored in the explanation text storage means 10 and the words stored in the word storage means 12, the semantic digitizing means 13 for each word included in the program content explanation text, Assign a numerical value suitable for the semantic content of each word.
[0049]
The word meaning storage means 14 stores the numerical value assigned to each word by the semantic digitizing means 13 together with the word. As the word meaning storage means 14, for example, a semiconductor memory, a magnetic disk, or the like can be used.
[0050]
In the present embodiment, a semantic vector value is used as a numerical value representing the semantic content of a word, and the semantic digitizing means 13 obtains the semantic vector value of each word extracted from the corresponding program content explanation text for each program. And a semantic vector value is assigned to each word.
[0051]
In the present embodiment, the semantic vector values are classified into, for example, 266 concepts, and one dimension is assigned to each concept. FIG. 14 shows only a part of an example of types included in one classified concept in the present embodiment.
[0052]
When the meaning of a word is expressed using a semantic vector value, it is determined whether or not the meaning of the word is related to each of the concepts prepared in advance. The value 1 is assigned to. If there is no association, the value 0 is assigned to the concept.
[0053]
For example, if the concepts are classified into six categories {human, sad, art, science, excitement, politics}, the word “pilot” is judged for the relevance for each classified concept and “0”. Alternatively, a value of “1” is set. The word “pilot” is set to “1” as being related to the concept “human”, and “0” is set to be not related to the concept “sad”. Similarly, “0” is set without relation to the concept “art”, “1” is set as being related to the concept “science”, and related to the concept “excitement”. “1” is set as the thing, and “0” is set as the thing that is not related to the concept “politics”. As a result, the semantic vector value of the word “pilot” is (1, 0, 0, 1, 1, 0).
[0054]
Note that the number and type of concept classifications that make up the semantic vector value need not be the same as in this embodiment, and can be increased or decreased as necessary. In the present embodiment, semantic vector values are fixedly assigned to basic words in advance, and the assigned semantic vector values are not changed. On the other hand, when a word other than the basic word to which the semantic vector value is assigned is extracted from the program description text, the semantic vector value of the word is dynamically set as follows. .
[0055]
In this case, first, a basic word to which a semantic vector value is fixedly assigned in advance is further extracted from each word extracted from the program description text for the program, and the extracted basic word The semantic vector value of the program is obtained using only the semantic vector value. In this case, for each basic word, the number of appearances (variable X3) included in the program description text and the value obtained by dividing all programs by the number of programs in which the basic word appears (variable X1). (Variable X2) is obtained. Then, by multiplying the obtained variable X3 and X2 by the basic word semantic vector value, the semantic vector value V1 of the word is obtained. Such a process is performed for all the basic words included in the program description document, and when the semantic vector value is obtained for each basic word, the obtained semantic vector value has already been calculated. Sequentially added to the semantic vector value.
[0056]
In this way, the meanings of the basic words in the program description text including all the basic words are added by adding the semantic vector values of all the basic words included in the program description text. The sum of vector values is obtained. This sum is a temporary semantic vector value for the program.
[0057]
Next, the meaning vector value of each word is obtained for all the words other than the basic word included in the program content explanation text in which the sum of the meaning vector values of the basic words is obtained in this way. . In this case, for each program that includes a word other than the basic word for all the words other than the basic word, the temporary semantic vector value of the program includes a word other than the basic word included in the program. The value multiplied by the number of is added.
[0058]
The sum of the semantic vector values thus obtained is used as a semantic vector value for a word other than the basic word, and stored in the word meaning storage means 14.
[0059]
The word meaning storage means 14 stores a numerical value representing the meaning content of each word assigned by the meaning digitizing means 13 together with the word.
[0060]
FIG. 15 is a diagram for explaining an example of a method of storing the meaning of a word by the word meaning storage unit 14. Here, the meaning vector value of the word calculated | required as mentioned above is stored with the notation of the word about one word.
[0061]
FIG. 3 is a block diagram for explaining the feature extraction unit 3 in the present embodiment in more detail. The feature extraction means 3 includes a word meaning acquisition means 16 for reading a numerical value representing the meaning content assigned to each word stored together with the program ID in the word meaning storage means 14 provided in the storage means 2, and a word meaning acquisition. Based on the numerical value representing the semantic content assigned to each word read by the means 16, the sentence meaning obtaining means 17 for obtaining the numerical value representing the semantic content of the entire sentence and the explanatory sentence storage means 10 are stored. The program content explanation text and the word extraction means 19 for extracting the program content explanation text included in the program information selected by the program designating means 4 for each word. This word extracting means 19 is the same means as the word extracting means 11 described above.
[0062]
The sentence meaning acquisition means 17 uses the above-described meaning vector value as a numerical value representing the meaning content of the word. Further, for each word constituting the sentence, the result of adding all the semantic vector values representing the meaning of the word is defined as a semantic vector value representing the semantic content of the entire sentence. In addition, in this embodiment, after adding vector values, normalization is performed because it is difficult to compare vector values by simply adding vector values. The magnitudes of the vectors are set to a constant value (for example, 10). This constant value is set to an appropriate value as necessary.
[0063]
A numerical value representing the meaning of the whole sentence obtained by the sentence meaning obtaining unit 17 is given to the sentence feature extracting unit 18. The sentence feature extracting unit 18 synthesizes a numerical value representing the semantic content of the entire sentence obtained by the sentence meaning obtaining unit 17 and a numerical value representing a meaning based on genre information obtained by the genre information adding unit 20 to be described later. Find the overall characteristics. The feature of the whole sentence obtained by the sentence feature extracting unit 18 is used by the program searching unit 7.
[0064]
The sentence feature extraction means 18 uses the above-described meaning vector value as a numerical value representing the meaning content. Both the numerical value representing the semantic content of the entire sentence and the numerical value representing the meaning based on the genre information are expressed as semantic vector values. To do. Therefore, the feature of the whole sentence is also expressed using a semantic vector value in the same way as the numerical value indicating the semantic content.
[0065]
The genre information adding means 20 obtains a numerical value representing the meaning based on the genre information based on the genre information for each program held in the genre information storage means 15 and the genre information of the program selected by the program specifying means 4. Add to the features of.
[0066]
FIG. 4 is a block diagram for explaining the genre information adding means 20 in this embodiment in more detail.
[0067]
The genre weight calculation means 21 dynamically obtains a weight value for each genre using the genre information for each program held in the genre information storage means 15.
[0068]
The genre-specific weight calculation means 21 dynamically determines the genre-specific weight value depending on how much the program belonging to a certain genre occupies with respect to all the programs included in the database. For example, when the ratio of programs belonging to a certain genre is large with respect to all the programs included in the database, the value of the weight of the genre is decreased. On the other hand, when the ratio of a program belonging to a certain genre is small with respect to all programs, the value of the weight of the genre is increased.
[0069]
For example, a program whose genre belongs to “sports” generally has a larger percentage of all programs than a program whose genre belongs to “baseball”. For this reason, when searching for a program related to a program belonging to the genre “sports”, even if the weight of a program belonging to the same “sports” is increased and prioritized at the time of search, the number of cases only increases. Makes it difficult to search for the program that they want. On the other hand, when seeking a program related to a program that belongs to the baseball category, increase the number of programs by increasing the weight of the program belonging to the same baseball and giving priority to the search. The program that the user wants can be obtained.
[0070]
In this embodiment, the weight value for each genre is represented by a real value from 0 to 1. Then, as a weight value of a program belonging to a certain genre, a ratio of the genre to all programs is used as it is. However, it is not necessary to use the same method as that of the present embodiment as a method for obtaining the weight for each genre. For example, it is possible to use a method of preparing a conversion table in which the proportion of programs belonging to each genre occupies all the programs and the weight value, and obtaining the weight for each genre using the conversion table.
[0071]
The genre-specific dynamic weight value obtained by the genre-specific weight calculation means 21 is provided to the genre-specific weight addition means 22. The genre weight addition means 22 is selected by the program designating means 4 and all programs held in the genre information storage means 15 based on the dynamic weight values for each genre obtained by the genre weight calculation means 21. A weight value based on genre information is obtained for a program. The obtained value is given to the weight synthesis means 23.
[0072]
The weight synthesizing means 23 is a weight value based on the genre information for each program obtained by the genre-specific weight addition means 22 and a genre-specific fixed weight value held by the genre-specific fixed weight storage means 24. And synthesize. The weight value synthesized by the weight synthesizing unit 23 is used by the sentence feature extracting unit 18.
[0073]
In the present embodiment, the above-described semantic vector value is used as a value representing a fixed weight for each genre. As the meaning vector value, a value representing the meaning of each genre is set in advance. Further, the magnitude of the semantic vector value is not normalized, and the magnitude of the vector value is used as a genre weight indicating the feature of the program.
[0074]
The weight synthesis means 23 multiplies the semantic vector value representing the fixed weight for each genre by the real value representing the weight for each genre obtained based on the genre information for each program, and synthesizes the resulting semantic vector value. The weight value is output to the sentence feature extraction means 18.
[0075]
The genre-specific fixed weight storage means 24 stores a weight value fixed for each genre. For example, a semiconductor memory or a magnetic disk can be used as the storage means. The weight value fixed by genre stored in the genre-specific fixed weight storage unit 24 is used by the weight synthesis unit 23.
[0076]
By setting a fixed weight value for each genre, for example, if there are genre types that are less relevant to the program content, such as “rebroadcast” or “other”, the programs belonging to those genres are Regardless of the contents of the program, it can be prevented that the program is searched as a similar program related to the designated program.
[0077]
FIG. 17 is a diagram for explaining an example of a method for storing the weight value fixed for each genre by the genre-specific fixed weight storage unit 24. One semantic vector value is stored in association with one genre type.
[0078]
FIG. 5 is a block diagram for explaining the program search means 7 in this embodiment in more detail. The program search means 7 has a similarity acquisition means 25. The similarity obtaining unit 25 obtains the similarity between programs using the similarity calculating unit 6 based on the characteristics of the plurality of programs obtained by the feature extracting unit 5.
[0079]
In the present embodiment, the above-described semantic vector value is used as a numerical value representing semantic content, and the semantic vector value is also used for the characteristics of a plurality of programs in the same manner as the numerical value representing semantic content. The similarity calculation means 6 obtains the similarity between the features of two programs as the distance between the semantic vector values. The distance between vectors can generally be expressed as an inner product value of vectors. Also in this embodiment, the similarity between the features of two programs is obtained as an inner product value of semantic vector values. When obtaining the similarity between the features between two programs, the similarity acquisition means 25 normalizes the magnitudes of the vector values of both in advance and aligns them to 10, for example. Then, the similarity acquisition unit 25 uses the similarity calculation unit 6 to obtain the inner product value of the semantic vector values. For example, the inner product value is 10 when the similarity is maximum, and the inner product value is 0 when the similarity is minimum.
[0080]
The similarity obtained by the similarity obtaining unit 25 is given to the similarity order arranging unit 26.
[0081]
The similarity order arranging unit 26 arranges the programs in the descending order of similarity with the designated program based on the similarity of the plurality of programs obtained by the similarity obtaining unit 25. The program information of the programs arranged in descending order by the similarity with the designated program by the similarity arranging unit 26 is provided to the search result display unit 8.
[0082]
FIG. 6 is a block diagram for explaining the program designating means 4 in the present embodiment in more detail. The program specifying means 4 has a plurality of program specifying means 27.
[0083]
In the multiple program designating unit 27, the user selects a plurality of programs with respect to the program information displayed by the program information display unit 3 according to an instruction from the related program display control unit 9.
The multiple program designating unit 27 is provided with a mechanical operation unit such as buttons and switches. As the multiple program designation means 27, a GUI or the like displayed together with the program information on the program information display means 3 can be used.
[0084]
The program weight designating unit 28 gives weights to the plurality of programs selected by the multiple program designating unit 27 according to an instruction from the related program display control unit 9. The program weight designation means 28 is provided with a mechanical operation unit such as buttons and switches. As the program weight designation means 28, a GUI displayed together with the program information on the program information display means 3 can be used.
[0085]
In the present embodiment, a method of assigning rankings of program weights in the order in which the user designates the programs is adopted. In this case, the program first designated by the user has the highest weight, and thereafter, the weight decreases as the program is designated. As a result, it is possible to intuitively search for programs having high relevance to a plurality of programs in a desired order without forcing the user to perform complicated operations. However, it is not necessary to use the same method as in this embodiment as a method for specifying the weights of a plurality of programs. For example, a method in which the user sets several levels of weight each time a program is designated can be used.
[0086]
The output of the program weight designation means 28 is given to the feature synthesis means 29. The feature synthesizing unit 29 gives the weight obtained by the program weight designating unit 28 to the program information of the plurality of programs selected by the multiple program designating unit 27. The program information of a plurality of programs weighted by the feature synthesis unit 29 is output to the feature extraction unit 5.
[0087]
In this embodiment, the weight of the program is represented by a real value from 0 to 1, and is output to the feature extraction means 5 together with the program information of the selected program. The feature extraction means 5 obtains the feature of each program as a semantic vector value using the program information of each selected program, and then selects the selected program by multiplying the semantic vector value representing the feature by the weight of the program. A feature representing the whole of a plurality of programs is obtained.
[0088]
Below, the entire process flow in the program information display apparatus of this embodiment having such a configuration will be described in order.
[0089]
In the present embodiment, the processing flow can be roughly divided into two. One is a program guide update process executed when EPG data is newly acquired. This process is a process executed using the program guide acquisition unit 1, the storage unit 2, and the program information display unit 3 as indicated by a broken line a in FIG. Details of the program guide update processing will be described later with reference to the flowchart of FIG.
[0090]
The other is related program display control processing that is executed when the user instructs to search for related programs. This process is a process executed by controlling the program specifying means 4, the feature extracting means 5, the program searching means 7 and the search result displaying means 8 by the related program displaying means 9 as indicated by a broken line b in FIG. is there. Details of the related program display control processing will be described later with reference to the flowchart of FIG.
[0091]
FIG. 7 is a diagram for explaining the flow of the program guide update process.
[0092]
First, in step S701, new EPG data is acquired by the program guide acquisition means 1. The content of the acquired EPG data is output to the storage means 2. Next, in step S702, the database stored in the storage unit 2 is updated using the acquired EPG data. The updated description text of the program content is output to the program information display means 3. Details of the process in step S702 will be described later with reference to the flowchart of FIG. Next, in step S703, the program information display means 3 displays an explanatory text of the program content based on the updated database content.
[0093]
FIG. 9 is a flowchart for explaining the flow of database update processing in step S702.
[0094]
First, in step S <b> 901, of the program information obtained from the program guide acquisition unit 1, the explanation text of the program content is stored in the explanation text storage unit 10. Next, in step S <b> 902, genre information representing the genre of the program among the program information obtained from the program guide acquisition unit 1 is stored in the genre information storage unit 15. Next, in step S903, the word extraction means 11 decomposes the explanatory text of the program content obtained from the explanatory text storage means 10 into words constituting the text. Then, the decomposed word is stored in the word storage means 12. Next, in step S904, the meaning digitizing means 13 digitizes the meaning of the word using the explanation text of the program content obtained from the explanatory text storage means 10 and each word obtained from the word storage means 12. Find the semantic vector value. The meaning of the digitized word is stored in the word meaning storage means 14.
[0095]
FIG. 18 is a flowchart for explaining the flow of processing for obtaining the semantic vector value of a word in step S904. This process is for obtaining the semantic vector values of all the words constituting the content explanation text of all the programs included in the database.
[0096]
First, in step S1801, all the elements of the vector V1 are initialized to 0. The vector V1 is used to temporarily store a numerical value in order to obtain the semantic vector value of the program currently focused on from the basic word semantic vector value.
Steps S1801 to S1808 are repetitive processes in units of programs, and the program of interest is advanced in order for all programs included in the database 2. In step S1802, the number of programs including the currently focused word (basic word) is recorded in the variable X1.
[0097]
Steps S1802 to S1806 are repetitive processes in units of words, and the focused word is advanced in order for all the words. In step S1803, a value obtained by dividing the total number of programs included in the database 2 by the variable X1 is recorded in the variable X2. Next, in step S1804, the number of words including the currently focused word in the currently focused program is recorded in the variable X3.
[0098]
Next, in step S1805, the semantic vector value of the currently focused word is multiplied by the variable X2 and the variable X3, and the result is added to the vector V1. Next, in step S1806, it is determined whether the processing from step S1802 to step S1806 has been performed for all words. If there is a word that has not been processed yet, the process returns to step S1802, and if all the words have been processed, the process proceeds to step S1807.
[0099]
In step S1807, the value of the vector V1 is set as the vector value of the currently focused program. Next, in step S1808, it is determined whether or not the processing from step S1801 to step S1808 has been performed for all programs. If there is a program that has not been processed yet, the process returns to step S1801, and if all the programs have been processed, the process proceeds to step S1809.
[0100]
In step S1809, all elements of the vector V2 are initialized to 0. The vector V2 is used to temporarily store a numerical value in order to obtain a semantic vector value of a currently focused word (a word other than a basic word) from the semantic vector value of the program obtained as described above. Used. Steps S1809 to S1814 are repetitive processes in units of words, and the process proceeds in order for all words.
[0101]
Next, in step S1810, the number of words including the currently focused word in the currently focused program is recorded in the variable X3. Steps S1810 to S1812 are repetitive processes in units of programs, and the process is sequentially performed for all programs included in the database 2. In step S1811, the semantic vector value of the currently focused program is multiplied by the variable X3, and the result is added to the vector V2.
[0102]
Next, in step S1812, it is determined whether or not the processing from step S1810 to step S1812 has been performed for all programs. If there is a program that has not been processed yet, the process returns to step S1810, and if all the programs have been processed, the process proceeds to step S1813. In step S1813, the value of the vector V2 is set as the vector value of the currently focused word. Next, in step S1814, it is determined whether or not the processing from step S1809 to step S1814 has been performed for all words. If there is a word that has not been processed yet, the process returns to step S1809. If all the words have been processed, the process ends.
[0103]
Through the above processing, the semantic vector value of a word can be obtained as a value corresponding to how the word is distributed (biased) with respect to all programs. For example, the meaning vector value of a word included in the program content explanation text of a large number of programs is not greatly influenced by the meaning content of the program. On the other hand, for words that are concentrated in a specific program, the semantic vector value is strongly influenced by the semantic content of the program. Therefore, in describing the program content, the meaning of each word used can be reflected as the value of the semantic vector.
[0104]
FIG. 8 is a flowchart for explaining the flow of the related program display control process.
[0105]
First, in step S801, program information of a program designated by the user is output from the program display means 3 to the program designation means 4. The program information of the program specified by the program specifying means 4 is output to the feature extracting means 5. At this time, when a plurality of programs are designated by the plurality of program designation means 27 and a weight is designated by the program weight designation means 28, a plurality of program information weighted by the feature synthesis means 29 is output to the feature extraction means 5. The Next, in step S802, the feature extraction unit 5 obtains the feature of the program using the program information of the program designated by the user. The obtained feature of the program designated by the user is output to the program search means 7. A detailed description of the processing in step S802 will be described later with reference to the flowchart of FIG.
[0106]
Next, in step S803, the feature extraction unit 5 obtains features for all programs stored in the storage unit 2 as a database using the program information. The characteristics of all programs included in the database are output to the program search means 7. A detailed description of the processing in step S803 will be described later with reference to the flowchart of FIG.
[0107]
Next, in step S804, the program search means 7 is included in the database using the characteristics of all programs obtained in step S803 based on the characteristics of the program specified by the user obtained in step S802. All programs are arranged in descending order of similarity to the designated program. The result of arranging the programs is output to the search result display means 8. A detailed description of the processing in step S804 will be described later with reference to the flowchart of FIG.
[0108]
Next, in step S805, the search result display means 8 displays the programs arranged in descending order of similarity with the designated program as related programs in order from the top. The related program display control means 9 controls the above processing.
[0109]
FIG. 10 is a flowchart for explaining the flow of processing for obtaining the program characteristics in steps S802 and S803.
[0110]
First, in step S1001, the word extraction means 19 decomposes the program content explanation text of the program designated by the user into words constituting the text. Next, in step S1002, the word meaning obtaining unit 16 obtains the meaning as a semantic vector value from the decomposed word, and the sentence meaning obtaining unit 17 further obtains the meaning of the entire explanation sentence from the meaning vector value of each word. Obtained as a semantic vector value. Next, in step S1003, the genre-specific weight calculation unit 21 dynamically determines the genre-specific weight value based on the genre information for each program obtained from the genre information storage unit 15. Next, in step S1004, the genre-specific weight addition means 22 determines the genre-specific weight for the program. Then, the weight synthesizing unit 23 obtains a weight obtained by synthesizing the fixed weight value held in the genre-specific fixed weight storage unit 24 with the obtained genre-specific weight value. In step S1005, the feature of the program is obtained as a semantic vector value by using the meaning of the program description text and the synthesized genre weight by the sentence feature extracting unit 18.
[0111]
FIG. 11 is a flowchart for explaining the flow of the program search process in step S804.
[0112]
First, in step 1101, the feature extraction means 5 uses the similarity calculation means 6 to compare the feature of the program designated by the user with the feature of the program included in the database based on the semantic vector value. The similarity between them is obtained as an inner product value of semantic vector values. In step S1102, the process in step S1101 is repeated for all programs included in the database. Next, in step S1103, the similarity order arranging unit 26 arranges all the programs included in the database in the descending order of similarity with the program designated by the user.
[0113]
【The invention's effect】
As described above in detail, according to the present invention, a user can specify a similar program highly relevant to a specified program simply by specifying an arbitrary program from program information displayed on the program display means. , EPG data can be retrieved and displayed. This operation does not require any prior knowledge about program search, and is simple, intuitive, and easy to understand. Therefore, considering that the main users of the present invention are ordinary people who do not have special knowledge, the present invention is very easy to use.
[0114]
Furthermore, according to the present invention, the user simply designates a plurality of arbitrary programs from the program information displayed on the program display means, and the relevance having characteristics common to the designated plurality of programs is obtained. Highly similar programs can be retrieved from the EPG data. This operation is simple, intuitive, and easy to understand as the user does not need any prior knowledge about program search as compared with the conventional search method that specifies a plurality of search conditions. Therefore, considering that the main users of the present invention are ordinary people who do not have special knowledge, the present invention is very easy to use.
[0115]
Furthermore, according to the present invention, when a user designates a plurality of arbitrary programs, it is possible to designate an order between the programs. As a result, it is possible to search EPG data for highly related similar programs having characteristics common to a plurality of designated programs in a form in which the order is reflected as a weight difference with respect to the characteristics of each program. This operation is simple, intuitive, and easy to understand as the user does not need any prior knowledge about program search as compared with the conventional search method that specifies a plurality of search conditions. Therefore, considering that the main users of the present invention are ordinary people who do not have special knowledge, the present invention is very easy to use.
[0116]
Furthermore, according to the present invention, in order to search for a similar program highly relevant to the program specified by the user, the feature of the program can be obtained based on the contents of the acquired EPG data. As a result, even if there is a change in the contents of the program due to the passage of years or the season, it is possible to always obtain appropriate program characteristics in response to the change dynamically.
[0117]
Furthermore, according to the present invention, the genre information to which the program belongs can be used together with the explanatory text of the program content in order to obtain the feature of the program. Conventionally, it has been necessary to individually specify search conditions for such different program attribute information. However, in the present invention, these can be handled as unified features by quantifying them. Thereby, it is possible to have a single reference for comparing the degree of similarity between a plurality of programs, and it is possible to display programs that are highly related in order from the highest degree of similarity to the user.
[0118]
Furthermore, according to the present invention, for each genre to which a program belongs, the strength of influence of the genre information on the program is held as a fixed weight, and this fixed weight can be used. Accordingly, even when the genre classification method itself is not always appropriate, it is possible to prevent an adverse effect from occurring when searching for similar programs with high relevance.
[0119]
Furthermore, according to the present invention, for each genre to which a program belongs, the strength of the influence of the genre information on the program can be dynamically corrected based on the content of the acquired EPG data and obtained as a weight. it can. As a result, even if there is a change in the contents of the program due to the passage of years or the season, it is possible to always obtain appropriate program characteristics in response to the change dynamically.
[Brief description of the drawings]
FIG. 1 is a block diagram for explaining a configuration of a program information display apparatus according to an embodiment of the present invention.
FIG. 2 is a block diagram for explaining a more detailed configuration of a database in the program information display apparatus according to the embodiment of the present invention.
FIG. 3 is a block diagram for explaining a more detailed configuration of feature extraction means in the program information display device according to the embodiment of the present invention.
FIG. 4 is a block diagram for explaining a more detailed configuration of genre information adding means in the program information display device according to one embodiment of the present invention.
FIG. 5 is a block diagram for explaining a more detailed configuration of program search means in the program information display device according to one embodiment of the present invention.
FIG. 6 is a block diagram for explaining a more detailed configuration of program specifying means in the program information display device according to the embodiment of the present invention.
FIG. 7 is a flowchart for explaining a flow of a program guide update process in the program information display device according to the embodiment of the present invention.
FIG. 8 is a flowchart for explaining the flow of related program display control processing in the program information display device according to the embodiment of the present invention;
FIG. 9 is a flowchart for explaining the flow of database update processing in the program information display device according to the embodiment of the present invention;
FIG. 10 is a flowchart for explaining a flow of processing for obtaining a feature of a program in the program information display device according to the embodiment of the present invention.
FIG. 11 is a flowchart for explaining a flow of program search processing in the program information display device according to the embodiment of the present invention.
FIG. 12 is a diagram showing an example of an explanatory text storage method by explanatory text storage means in the program information display device according to the embodiment of the present invention;
FIG. 13 is a diagram showing an example of a word storage method by word storage means in the program information display device according to one embodiment of the present invention.
FIG. 14 is a diagram showing an example (partial) of concept classification types in the program information display device according to the embodiment of the present invention.
FIG. 15 is a diagram illustrating an example of a method of storing meanings of words by a word storage unit in the program information display device according to the embodiment of the present invention.
FIG. 16 is a diagram showing an example of a genre type storage method for each program by a genre information storage unit in the program information display device according to the embodiment of the present invention;
FIG. 17 is a diagram illustrating an example of a weight value storage method fixed by genre by a genre-specific fixed weight storage unit in the program information display device according to the embodiment of the present invention;
FIG. 18 is a flowchart for explaining a flow of processing for obtaining a semantic vector value of a word in the program information display device according to the embodiment of the present invention.
[Explanation of symbols]
1 Program guide acquisition means
2 storage means
3 Program information display means
4 Program designation means
5 Feature extraction means
6 Similarity calculation means
7 Program search means
8 Search result display means
9 Related program display control means
10. Explanation text storage means
11 Word extraction means
12 Word storage means
13 Meaning digitization means
14 Word meaning storage means
15 Genre information storage means
16 Word meaning acquisition means
17 Meaning acquisition means
18 Text feature extraction means
19 Word extraction means
20 Genre information adding means
21 Weight calculation means by genre
22 Weight addition means by genre
23 Weight synthesis means
24 Fixed weight storage means by genre
25 Similarity acquisition means
26 Similarity order arrangement means
27 Multiple program designation means
28 Program weight designation means
29 Features Administrative measures

Claims (9)

A program guide acquisition means for acquiring electronic program guide data distributed together with a broadcast wave of a television program or distributed via a network;
Storage means for storing the acquired electronic program guide data as a database;
Program information display means for displaying program information included in the database;
Program designation means for designating an arbitrary program by the user from the program information displayed on the program information display means;
Feature extraction means for extracting program features from program information contained in the database;
Similarity calculation means for comparing features of a plurality of programs to obtain a similarity between programs;
Based on the characteristics of the plurality of programs obtained by the feature extraction means, the similarity calculation means is used to compare the similarities between programs, and similar programs related to the program designated by the program designation means are obtained. A program search means for searching;
Search result display means for displaying the search results;
The program information display means, the program designation means, the program search means and the search result display means are controlled to display program information by the program information display means, and the program designated by the user by the program designation means and a high degree of similarity programs, by searching by該番set search means, the search results possess a related program display control means for displaying on the search result display means,
The feature of the program extracted by the feature extraction means is that of the description text of the program content obtained by calculating the sum of the semantic vector values respectively corresponding to the words included in the description text of the program content included in the program information. A semantic vector value obtained by normalizing a semantic vector value,
The semantic vector value is a multidimensional vector value in which one dimension is assigned to each concept,
The storage means
Explanatory text storage means for storing the explanatory text of the program content explaining the program content included in the program information and the program ID of the corresponding program in pairs;
Word storage means for storing a set of a word constituting an explanatory sentence of each program content and a list of program IDs of programs corresponding to the explanatory text of the program content including the word;
Program information display device characterized by chromatic and word meaning storage means for storing semantic vector value representing the meaning of a word with the word. The storage means
And the word extraction means for decomposing a description sentence turn set content into words,
Based on the distribution of words in the description text of all program contents stored in said description storage means, means digitizing executing the digitizing of the meaning of the word by the meaning of individual words represented by the mean vector value Means and
Program information display device according to claim 1, characterized in that it comprises a. The feature extraction means includes
Word extraction means for decomposing an explanation sentence of program content included in program information into words;
Word meaning acquisition means for performing a process of reading from the database a meaning vector value obtained by quantifying the meaning of a word in the word meaning storage means of the storage means as digitization of the meaning of each decomposed word;
A semantic vector value representing the meaning of the explanatory text of the program content made up of individual words by calculation on the semantic vector value corresponding to the meaning of each word , obtained by quantification in the semantic digitizing means of the storage means A sentence meaning acquisition means for
3. The program information display apparatus according to claim 1, further comprising a sentence feature extraction unit that extracts a numerical description of the meaning of the explanation text of the program content as a sentence feature. The program search means includes
Similarity acquisition means for determining the similarity between programs based on the characteristics of a plurality of programs;
4. The program information display device according to claim 1, further comprising similarity order arranging means for arranging a plurality of programs in descending order of similarity. The storage means includes genre information storage means for storing genre information of programs,
5. The feature extracting means comprises genre information adding means for assigning the semantic vector to genre information of a program and digitizing the genre information of the program and adding it to the feature of the program. The program information display device according to any one of the above. The genre information adding means is
A genre-specific fixed weight storage unit that stores a fixed weight value for each genre; and a genre-specific weight addition unit that obtains a weight based on genre information for each program using the fixed weight value. The program information display apparatus according to claim 5, wherein: The genre information adding means is
A genre-specific weight calculation means for calculating a dynamic weight value for each genre using a ratio of the number of programs belonging to the genre with respect to the total number of programs included in the database;
6. The program information display device according to claim 5, further comprising a genre-specific weight addition means for obtaining a weight based on genre information for each program using the calculated weight value. The genre information adding means includes weight combining means for combining a fixed weight value for each genre and a dynamic weight value calculated using a database.
8. The program information display device according to claim 6, wherein genre information weighted by using the synthesized weight value is added to the feature of the program. The program specifying means includes
A plurality of program specifying means for a user to specify a plurality of arbitrary programs from the program information displayed on the program information display means;
A feature synthesis means for weighting program information of a plurality of designated programs ;
It possesses the program weighting designation means for the user to specify a weighting of more of any of the program,
9. The program information display device according to claim 1 , wherein the feature synthesis unit adds a designated weight to program information of a plurality of designated programs.

Images