JP2013021661A - Viewing rate estimation device and viewing rate estimation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a viewing rate estimation device and a viewing rate estimation method capable of speedily and accurately estimating the viewing rate of a television program.SOLUTION: The viewing rate estimation device 2 comprises: a storage part 204 which previously stores a correlation table between the number of user-generated contents on a television program generated on a network 3 by the viewers of the television program and the viewing rate of the television program; an object content number acquisition part 205 which acquires the number of user-generated contents on an object television program; and a viewing rate estimation part 206 which estimates the viewing rate of the television program according to the number of the user-generated contents on the object television program acquired by the object content number acquisition part 205 and the correlation table previously stored by the storage part 204.

Description

  The present invention relates to an audience rating estimation apparatus and an audience rating estimation method for estimating an audience rating of a television program.

  Conventional TV program ratings are mainly household ratings provided by one company. The household audience rating is measured for each viewing area. In the Kanto, Kansai, and Nagoya areas, 600 households in the viewing area are randomly selected and measured using automatic measuring instruments. There are 200 households outside the above areas. The audience rating is recorded in minutes, but the summation is once a day in the early morning. The announcement of household audience rating is the next business day.

  The method of measuring household audience rating is also affected by the number of TVs in each household. For example, if the household M has two televisions and the broadcasting station A is selected for both televisions, it is counted that the broadcasting station A has been viewed. Only a state in which the broadcasting station A is selected for one or a plurality of televisions is counted. Further, for example, when there are two TVs in the household N and the broadcasting station B and the broadcasting station C are selected, the broadcasting station B and the broadcasting station C are counted, respectively. In this state, broadcast stations A, B, and C are calculated as the same household audience rating.

  On the other hand, as an apparatus for calculating the audience rating of a television program, for example, Patent Literature 1 below discloses an audience activity rate calculating apparatus. The viewing behavior rate calculating device counts only the users who have searched for keywords related to the contents of the television program among the viewers of the television program as viewers of the television program, and calculates the audience rating of the television program.

JP 2010-34674 A

  However, the conventional household audience rating is calculated on the next business day after the end of the television program, and there is a time lag between the broadcast of the television program and the result of the household audience rating. In addition, the viewing behavior rate calculating apparatus is actually watching a television program, but a user who does not search for a keyword related to the content of the television program is not counted as a viewer, and an accurate television program viewing rate cannot be calculated.

  Therefore, the present invention has been made in view of such problems, and an object of the present invention is to provide an audience rating estimation apparatus and an audience rating estimation method that can quickly and accurately estimate an audience rating of a television program. To do.

  In order to solve the above problems, the audience rating estimation apparatus of the present invention provides a correlation table between the number of user-generated contents related to the television program generated on the network by the viewer of the television program and the audience rating of the television program. Storage means for storing in advance, target content number acquisition means for acquiring the number of user-generated content related to the target TV program, number of user-generated content related to the target TV program acquired by the target content number acquisition means, and storage means And audience rating estimation means for estimating the audience rating of the television program based on the correlation table stored in advance.

  The audience rating estimation method of the present invention is an audience rating estimation method by an audience rating estimation device, in which the storage means of the audience rating estimation device is a user related to the television program generated on the network by the viewer of the television program. A storage step for storing in advance a correlation table between the number of generated contents and the audience rating of the TV program, and an object for acquiring the number of user-generated contents related to the target TV program Based on the number-of-contents acquisition step and the audience-rate estimation means of the audience rating estimation device based on the number of user-generated contents related to the target television program acquired in the target content number acquisition step and the correlation table stored in advance in the storage step And an audience rating estimation step for estimating the audience rating of the television program.

  According to such an audience rating estimation apparatus and audience rating estimation method, the correlation table between the number of user-generated contents related to the television program generated on the network by the viewer of the television program and the audience rating of the television program is obtained. Stored in advance. Then, the audience rating of the TV program is estimated based on the acquired number of user-generated contents related to the target TV program and a correlation table stored in advance. By adopting such a configuration, it is possible to quickly estimate the audience rating of the TV program by acquiring the number of user-generated contents related to the TV program being broadcast. For example, since the audience rating is estimated based on the correlation table between the number of past user-generated contents stored in advance and the audience rating, a more accurate audience rating of the television program can be estimated.

  In the audience rating estimation apparatus according to the present invention, the storage unit stores in advance a correlation table between the number of user-generated contents for each predetermined time and the audience rating for each predetermined time, and the target content number acquiring unit has a predetermined time. The audience rating estimation means obtains the number of user-generated contents at a predetermined time related to the target television program acquired by the target content number acquisition means, and a correlation table stored in advance by the storage means. It is preferable to estimate the audience rating of the television program at the predetermined time based on the above. By adopting such a configuration, it is possible to quickly estimate the audience rating of the TV program at the predetermined time by acquiring the number of user-generated contents for the predetermined time regarding the TV program being broadcast. Further, for example, since the audience rating is estimated based on the correlation table between the number of past many user-generated contents for each predetermined time stored in advance and the audience rating for each predetermined time, a more accurate television at the predetermined time is obtained. The audience rating of the program can be estimated.

  In addition, the audience rating estimation device of the present invention includes content number acquisition means for acquiring the number of user-generated content related to the TV program generated on the network by the viewer of the TV program, and the TV program calculated by the external device. Correlation for generating a correlation table between the audience rating acquisition means for acquiring the audience rating, the number of user-generated content related to the TV program acquired by the content number acquisition means, and the audience rating of the TV program acquired by the audience rating acquisition means It is preferable that the information processing apparatus further includes a table generation unit, and the storage unit stores the correlation table generated by the correlation table generation unit in advance. With this configuration, for example, a correlation table can be generated from the number of user-generated contents in the past and the audience rating, and the audience rating is estimated based on the correlation table. The audience rating can be estimated.

  In the audience rating estimation device according to the present invention, the correlation table generation unit corrects the number of user-generated contents acquired by the content number acquisition unit based on the characteristics of the user-generated content or the characteristics of the television program. Is preferred. By adopting such a configuration, it is possible to estimate a more accurate audience rating of a television program in consideration of the characteristics of user-generated content or the characteristics of a television program.

  In the audience rating estimation apparatus according to the present invention, the audience rating estimation unit corrects the number of user-generated contents acquired by the target content number acquisition unit based on the characteristics of the user-generated content or the characteristics of the television program. It is preferable. By adopting such a configuration, it is possible to estimate a more accurate audience rating of a television program in consideration of the characteristics of user-generated content or the characteristics of a television program.

  Further, the audience rating estimation device of the present invention corrects the correlation table stored in advance by the storage unit based on the difference between the audience rating estimated by the audience rating estimation unit and the audience rating calculated by the external device. And a table correction unit, and the storage unit preferably stores the correlation table corrected by the correlation table correction unit. By adopting such a configuration, it is possible to correct the correlation table with higher accuracy by feeding back the audience rating estimation result, and thereby to estimate the audience rating of the television program more accurately.

  In the audience rating estimation apparatus according to the present invention, the storage unit stores a correlation table in advance for each category for classifying the television programs, and the target content number acquisition unit generates user-related data for the target TV program for each category. The audience rating estimating means obtains the number of contents, the number of user-generated contents related to the target television program for each category acquired by the target content number acquiring means, and the correlation table for each category stored in advance by the storage means, Preferably, the audience rating of the television program for each category is estimated, and the audience rating of the television program is estimated based on the estimated audience rating of the television program for each category. If such a configuration is adopted, the audience rating is estimated based on the audience rating estimated for each category in classifying the television program, so that the more accurate audience rating of the television program can be estimated.

  In the audience rating estimation apparatus according to the present invention, the storage unit stores a correlation table in advance for each type of user-generated content related to a television program, and the target content number acquisition unit stores each type of user-generated content related to a target TV program. The number of user-generated contents is acquired, and the audience rating estimation means is stored in advance by the storage means and the number of user-generated contents for each type of user-generated content related to the target television program acquired by the target content number acquisition means. Based on the correlation table for each type of user-generated content regarding the TV program, the audience rating of the TV program for each type of user-generated content regarding the TV program is estimated, and the estimated user-generated content regarding the TV program is estimated. The audience rating of the TV program for each type and the TV number Based on the weight of each type of user-generated content related to estimate the audience of the TV program, it is preferable. With this configuration, since the audience rating is estimated based on the audience rating estimated for each type of user-generated content and the weighting for each type of user-generated content, it is possible to estimate the audience rating of a more accurate television program. it can.

  The audience rating estimation apparatus according to the present invention includes an analysis unit that analyzes the content of user-generated content used by the audience rating estimation unit to estimate the audience rating, an audience rating estimated by the audience rating estimation unit, and an analysis unit. It is preferable to further comprise output means for outputting the analysis result according to the above. By adopting such a configuration, the analysis result of the content of the user-generated content can be output together with the audience rating, so that the user can know the viewer's reaction in more detail.

  According to the present invention, the audience rating of a television program can be estimated quickly and more accurately.

It is a system configuration figure showing the system configuration of the audience rating estimation system concerning the embodiment of the present invention. It is a functional block diagram which shows the structure of the audience rating estimation apparatus which concerns on embodiment of this invention. It is a figure which shows the hardware constitutions of the audience rating estimation apparatus which concerns on embodiment of this invention. It is a figure which shows the example of the correlation figure of the number of tweets and audience rating in embodiment of this invention. It is a figure which shows the example of the correlation table in embodiment of this invention. It is a figure which shows the example of the correlation diagram of the household audience rating and the number of tweets for every broadcasting station in embodiment of this invention. It is a sequence diagram which shows the process which produces | generates a correlation table among the processes of the audience rating estimation method performed with the audience rating estimation system which concerns on embodiment of this invention. It is a sequence diagram which shows the process which estimates audience rating among the processes of the audience rating estimation method performed with the audience rating estimation system which concerns on embodiment of this invention. It is a sequence diagram which shows the process which estimates an audience rating from the audience rating and weighting for every kind of user-generated content among the processes of the audience rating estimation method performed with the audience rating estimation system which concerns on embodiment of this invention.

  Hereinafter, preferred embodiments of an audience rating estimation apparatus and an audience rating estimation method according to the present invention will be described in detail with reference to the drawings. In the description of the drawings, the same elements are denoted by the same reference numerals, and redundant description is omitted.

  FIG. 1 is a system configuration diagram showing a system configuration of an audience rating estimation system 1 of the present embodiment. As shown in FIG. 1, an audience rating estimation system 1 includes an audience rating estimation device 2, a network 3, an external device 4, a television 5 (television 5a, television 5b,...), A server 6 (server 6a, server 6b,...), And a personal computer 7 (personal computer 7a1, personal computer 7a2,..., Personal computer 7b1, personal computer 7b2,...).

  The audience rating estimation device 2, the external device 4, the server 6, and the personal computer 7 are configured by a computer device or the like, and the television 5 is configured by a television receiver or the like. The network 3 is a public network such as the Internet. The audience rating estimation device 2, the external device 4, the server 6, and the personal computer 7 can communicate with each other via the network 3. The external device 4 and the television 5 are connected via the network 3 or a dedicated network line.

  The external device 4 is a conventional audience rating calculation system, for example, a dedicated device for measuring the viewing situation is installed on a television 5 in a specific home, and the measurement result is transmitted from the dedicated device to the external device 4 via the network. In this system, the external device 4 aggregates the measurement results and calculates the audience rating of the television program. The external device 4 receives a transmission request for a predetermined television program at predetermined time intervals from the audience rating estimation device 2 via the network 3, and estimates the audience rating for the predetermined time of the requested predetermined television program. Transmit to device 2.

  The server 6 is provided with a server provided by an Internet service provider or the like, a short text posting site, a bulletin board, a blog, a video sharing site, an SNS (Social Networking Service) or the like that can post a short text called “tweet”. Server. The server 6 stores user-generated content (UGC) such as posted content created by a general user using a personal computer 7 or a portable terminal such as a mobile phone. In the present embodiment, the server 6 is provided with a short text posting site, and is a kind of user-generated content about the content of the viewed TV program that is generated and posted using the personal computer 7 by the viewer of the TV program. It is assumed that tweets are accumulated. In the server 6, tweets regarding the TV program by the viewer are accumulated in real time while the TV program is being broadcast.

  On short posting sites, posted tweets are given identifiers that indicate related topics. This is called a hash tag. Some hashtags indicate TV broadcast stations and TV programs. By using hashtags related to these TV programs, it is possible to identify tweets related to TV programs on the short text posting site.

  FIG. 2 is a functional block diagram showing the configuration of the audience rating estimation apparatus 2. As shown in FIG. 2, the audience rating estimation apparatus 2 includes a content number acquisition unit 201 (content number acquisition unit), an audience rating acquisition unit 202 (viewing rate acquisition unit), a correlation table generation unit 203 (correlation table generation unit), and a storage. Unit 204 (storage unit), target content number acquisition unit 205 (target content number acquisition unit), audience rating estimation unit 206 (viewing rate estimation unit), analysis unit 207 (analysis unit), output unit 208 (output unit), and A correlation table correction unit 209 (correlation table correction means) is included.

  The audience rating estimation device 2 is configured by hardware such as a CPU. FIG. 3 is a hardware configuration diagram of the audience rating estimation apparatus 2. As shown in FIG. 3, the audience rating estimation device 2 shown in FIG. 2 physically includes a CPU 250, a RAM 251 and a ROM 252, which are main storage devices, an input / output device 253 such as a keyboard and a display, a communication module 254, and The computer system includes an auxiliary storage device 255 and the like.

  The function of each functional block of the audience rating estimation device 2 shown in FIG. 2 is an input / output device under the control of the CPU 250 by reading predetermined computer software on the hardware such as the CPU 250 and the RAM 251 shown in FIG. This is realized by operating the H.253, the communication module 254, and the auxiliary storage device 255, and reading and writing data in the RAM 251.

  Hereinafter, each functional block of the audience rating estimation apparatus 2 will be described based on the functional blocks of the audience rating estimation apparatus 2 shown in FIG.

  The number-of-contents acquisition unit 201 acquires the number of user-generated contents per predetermined time related to the television program generated on the network by the viewer of the television program. Specifically, the number-of-contents acquisition unit 201 first acquires tweets generated on one or more servers 6 by a viewer of a television program via the network 3. The number-of-contents acquisition unit 201 identifies the TV program to be tweeted using the hash tag described above, divides the acquired tweet into each TV program to be tweeted, and the number of tweets related to each TV program. Are counted. The content number acquisition unit 201 counts the number of tweets at short intervals such as every predetermined time, for example, every minute in order to provide real-time properties. When the user-generated content is tweets, the number of tweets is an index to the TV program audience rating. The number-of-contents acquisition unit 201 outputs the acquired tweets and the number of tweets for each predetermined time to the correlation table generation unit 203.

  The audience rating acquisition unit 202 acquires the audience rating for each predetermined time of the television program calculated by the external device 4. Specifically, the audience rating acquisition unit 202 transmits an audience rating transmission request for each predetermined time of a predetermined television program to the external apparatus 4 via the network 3, and is transmitted from the external apparatus 4 accordingly. The audience rating is received via the network 3 and acquired. Since the television program audience rating broadcasted every day is counted every minute, for example, the audience rating acquisition unit 202 acquires the audience rating per minute of each television program. The audience rating acquisition unit 202 outputs the acquired audience rating for each predetermined time of each television program to the correlation table generation unit 203.

  The correlation table generation unit 203 calculates the number of user-generated contents per predetermined time related to the television program acquired by the content number acquisition unit 201 and the audience rating per predetermined time of the TV program acquired by the audience rating acquisition unit 202. Generate a correlation table. Specifically, the correlation table generation unit 203 firstly tweeted each number of TV programs output by the content number acquisition unit 201 and the number of tweets, and each TV program output by the audience rating acquisition unit 202. And the audience rating for each predetermined time. And the correlation table production | generation part 203 produces | generates the correlation table of the number of tweets for every predetermined time of each acquired television program, and an audience rating.

  As an example of how the correlation table generation unit 203 generates a correlation table, the correlation table generation unit 203 first acquires, for a specific television program, the number of tweets of the acquired television program per predetermined time on the horizontal axis. The correlation diagram shown in FIG. 4 is generated with the viewing rate of the television program for each predetermined time as the vertical axis. Next, the correlation table generation unit 203 generates a correlation table of the number of tweets and the audience rating for each predetermined time for each category described later for the television program based on the correlation diagram shown in FIG. For example, the correlation table generator 203 generates correlation tables a to c shown in FIG. In addition, the correlation table produced | generated is not limited to three, You may produce | generate one or more for every division. From the correlation table a shown in FIG. 5, for example, it can be seen that the number of tweets “100” for a predetermined time for the television program corresponds to the audience rating “11.0%”. In this way, the correlation table generation unit 203 generates, for example, a correlation table of the number of tweets aggregated every minute and the television program audience rating per minute. Thereby, the correlation between the number of tweets of each TV program and the audience rating is obtained. The correlation table generated by the correlation table generation unit 203 is a correlation table between past results, that is, the number of tweets related to a television program broadcast in the past and the audience rating.

  The correlation table generation unit 203 generates a correlation table for each category for classifying television programs. Examples of the classification include a broadcast station, a broadcast program genre, a broadcast time zone, and a broadcast area. The reason why the correlation table is generated for each category is that each category has the following characteristics.

  As a characteristic of a broadcasting station, there is a tendency that similar tweets are not generated for any broadcasting station. Since the program organization differs depending on the broadcasting station, it has good or bad compatibility with tweets. In addition, the variation in audience rating varies greatly from broadcast station to broadcast station. For example, FIG. 6 shows a correlation diagram between the number of tweets for each broadcasting station and the household audience rating. However, the station A and the station C have a different trend in audience rating even in the same broadcast time zone and the same broadcast program genre. .

  As a characteristic of the broadcast program genre, there is a tendency that the broadcast program genre tends to generate a tweet. For example, variety and sports tend to require a sense of unity with other viewers, so there are many tweets. In addition, information programs such as news tend to generate tweets depending on information, and almost no tweets are generated for welfare and theater. In addition, there is a big tendency for broadcast program genres depending on the time of day. For example, there are many information-related programs in the morning, drama rebroadcasts and information programs for housewives in the daytime, news in the evening and night, and a lot of variety in the golden time.

  One characteristic of the broadcast time zone is that tweets are not generated uniformly in any broadcast time zone. For example, the number of tweets increases from golden time to midnight, and the number of tweets during the day is quite small.

  As a characteristic of the broadcast area, the reaction of viewers varies from broadcast area to broadcast area for nationally broadcast television programs. In addition, popular program genres vary from region to region. For example, in Kansai, the number of tweets and the audience rating of variety programs are higher than in Kanto.

  The correlation table generation unit 203 may correct the number of user-generated contents acquired by the content number acquisition unit 201 based on the characteristics of the user-generated content or the characteristics of the television program.

As a characteristic of the user-generated content, for example, in the user-generated content, there is a possibility that negative content may accelerate and increase. Based on this point, specifically, the correlation table generation unit 203 does not simply count the number of tweets of the negative content among the tweets acquired from the content number acquisition unit 201 as the number of tweets. Whether or not the tweet is negative can be determined by determining whether the tweet contains a negative word registered in advance, and by recognizing the meaning and syntactic structure of the word, For example, the techniques disclosed in the following patent documents can be used.
Patent Document: Japanese Patent No. 3402598

  Further, the correlation table generation unit 203 measures the number of positive content tweets and the number of negative content tweets, and first counts the number of positive content tweets. Then, the correlation table generation unit 203 also measures the correlation regarding tweets of negative content, and counts the number of negative tweets based on the type and degree of the negative content, so that the accuracy of the audience rating is expected to be improved. If so, the correlation table generator 203 also counts the negative contents. For example, it may be effective depending on the genre of the TV program.

  In addition, as a characteristic of user-generated content, for example, when a current keyword such as “nuclear power” frequently appears, there is a possibility that the number of tweets related to a specific keyword may increase rapidly. Since the correlation table generation unit 203 takes into account the increase in the number of tweets due to frequent occurrence of current affairs keywords, for the keywords that are frequently handled in the TV program or have a long broadcast time, the number of tweets corresponding to that is calculated. Correct it by decreasing it.

  Specifically, the correlation table generation unit 203 stores in advance an average value (appearance rate) of the number of keywords that appear in tweets about a TV program, and a keyword whose frequency of appearance during a TV program broadcast increases rapidly at a certain time. Is determined to be a current affair keyword. For example, the correlation table generation unit 203 averages the number of keywords for each keyword category, for example, “performer” is about 50% of tweets, “product name” is about 10% of tweets, and “company name”. Assume that about 2% of tweets are stored in advance, and that “facility name” stores about 1% of tweets in advance. When keywords such as “nuclear power” and “A power company” increase, the number of keywords in which the ratio of the category “facility name” to which “nuclear power” belongs and the category “company name” to which “A power company” belongs are stored. Therefore, the correlation table generation unit 203 determines these as current affairs keywords.

  The correlation table generating unit 203 selects a current affair keyword when a portion including a current affair keyword in a tweet in a specific television program includes a threshold value, for example, twice or more of the normal average value. The number of tweets included is not counted. The threshold value defining method using this threshold value is a feedback process based on the difference between the audience rating estimated by the audience rating estimation unit 206 and the audience rating calculated by the external device 4 by the correlation table correction unit 209 described later. In this case, the correlation table correction unit 209 may perform optimization.

  In addition, as a characteristic of a TV program, for example, a special program such as a summary of past broadcast programs broadcast at the break of a broadcast cool or a re-broadcast of multiple episodes, a special program that is broadcast every season, etc. There is a point that there are many viewers' reactions different from normal broadcasting. The correlation table generation unit 203 corrects the number of tweets for each program in consideration of similar program information broadcast in the past for special programs and seasonal programs. For example, since the compilation program related to idols tends to have a high audience rating and the number of tweets is large, the correlation table generation unit 203 uses a predetermined correction coefficient to reduce the number of tweets. On the other hand, since the number of tweets does not increase as much as the audience rating in a past replay of a past drama, the correlation table generation unit 203 increases the number of tweets using a predetermined correction coefficient. Note that the correlation table generation unit 203 may generate a correlation table only for normal broadcasts excluding special programs such as special numbers.

  The storage unit 204 stores the correlation table generated by the correlation table generation unit 203 in advance in the auxiliary storage device 255 or the like. The memory | storage part 204 may memorize | store the correlation table produced | generated for every above-mentioned division in classifying a television program by the correlation table production | generation part 203. FIG. The storage unit 204 may store a correlation table corrected by a correlation table correction unit 209 described later.

  The target content number acquisition unit 205 acquires the number of user-generated content for a predetermined time related to the target television program. Specifically, the target content number acquisition unit 205 acquires user-generated content for a predetermined time related to the target television program from one or more servers 6. Note that the target television program and the predetermined time may be designated by the user of the audience rating estimation apparatus 2 or may be stored in the storage unit 204 in advance. Then, the target content number acquisition unit 205 counts the number of acquired user-generated contents and acquires the number of user-generated contents. Unlike the content number acquisition unit 201, the target content number acquisition unit 205 acquires the number of user-generated content in a predetermined time, for example, the number of user-generated content between “13:10” and “13:11”. To do. This predetermined time is a time for which audience rating is to be estimated. The target content number acquisition unit 205 outputs the user-generated content regarding the acquired target television program at a predetermined time to the audience rating estimation unit 206, the analysis unit 207, the output unit 208, and the correlation table correction unit 209.

  The target content number acquisition unit 205 may acquire the number of user-generated content for a predetermined time related to the target TV program for each of the above-described categories in classifying the TV programs. For example, when the broadcast station of the target television program is “A station”, the broadcast program genre is “variety”, and the predetermined time is between “13:10” and “13:11”, the target content number acquisition unit 205 obtains the number of tweets related to the TV program of “Station A” between “13:10” and “13:11” using the above hash tag and the like. The number of tweets related to a TV program having a broadcast program genre of “variety” between “10:11” and “13:11” is acquired.

  The audience rating estimation unit 206 determines the predetermined time based on the number of user-generated contents regarding the target television program acquired by the target content number acquisition unit 205 and the correlation table stored in advance by the storage unit 204. The audience rating of the TV program at is estimated.

  For example, there is only one correlation table stored by the storage unit 204, not for each category for classifying television programs, and the correlation table associates the number of tweets per minute with the audience rating. Suppose that Further, it is assumed that the number of tweets between “13:10” and “13:11” related to the target television program acquired by the target content number acquisition unit 205 is “100”. When the audience rating “11.0%” is associated with the number of tweets “100” in the correlation table, the audience rating estimation unit 206 starts from “13:10” regarding the target TV program. The audience rating of “13:11” is estimated to be “11.0%”.

  In addition, when there are a plurality of correlation tables stored in the storage unit 204 for each of the above-described categories for classifying a television program, the audience rating estimation unit 206 performs the classification for each category acquired by the target content number acquisition unit 205. Based on the number of user-generated contents related to the target TV program and the correlation table for each category stored in the storage unit 204, the audience rating of the TV program for each category is estimated. Specifically, the audience rating estimation unit 206 sets the above correlation table to 1 based on the number of tweets of the category acquired by the target content number acquisition unit 205 and the correlation table of the category for a certain category. In the same way as the audience rating estimation method in the case where there is only one, the audience rating of that section is estimated, and similarly the audience ratings are estimated for all sections. Then, the audience rating estimation unit 206 calculates an average value of the estimated audience ratings of each section, and estimates the calculation result as a final audience rating.

  The audience rating estimation unit 206 may correct the number of user-generated contents acquired by the target content number acquisition unit 205 based on the characteristics of the user-generated content or the characteristics of the television program. The method for correcting the number of user-generated contents by the audience rating estimation unit 206 is the same as the method for correcting the number of user-generated contents by the correlation table generating unit 203 described above. The audience rating estimation unit 206 does not correct the number of user-generated content when the user-generated content acquired by the target content number acquisition unit 205 is a normal broadcast program, and other than the normal broadcast program. In this case, the number of user-generated contents may be corrected.

  The analysis unit 207 analyzes the content of the user-generated content used when the audience rating estimation unit 206 estimates the audience rating. Specific examples of the analysis by the analysis unit 207 include separating whether the content of the tweet is positive or negative, and analyzing the distribution of the number of keywords included in the tweet. Can be mentioned.

  The output unit 208 outputs one or both of the audience rating estimated by the audience rating estimation unit 206 and the analysis result by the analysis unit 207. The output unit 208, for example, outputs both the audience rating estimated by the audience rating estimation unit 206 and the analysis result by the analysis unit 207, so that the user can take the correlation of the audience rating as well as the mere audience rating. It is possible to analyze the content of the tweet, and to grasp the viewer's reaction in detail, such as whether it is due to a positive reaction or a negative reaction. In addition, the user can obtain an index for estimating whether the audience rating has increased due to the handling of current affairs keywords, whether the audience rating has increased, what has attracted attention in the program, and the like. Furthermore, in the case of user-generated content in which the attribute of the viewer who posted the tweet is registered, the user can grasp the TV program audience rating and the reaction of the program content to the viewer of each segment. .

  The correlation table correction unit 209 corrects the correlation table stored in advance by the storage unit 204 based on the difference between the audience rating estimated by the audience rating estimation unit 206 and the audience rating calculated by the external device 4. Specifically, the correlation table correction unit 209 determines the audience rating estimated by the audience rating estimation unit 206 and the audience rating calculated by the external device 4 for a certain broadcast station, broadcast program genre, broadcast time zone, and broadcast area. The difference is recorded. The difference is ideally distributed symmetrically in the plus / minus direction around zero. When the audience rating error estimated by the audience rating estimation unit 206 increases, the center value deviates from 0. Therefore, the correlation table correction unit 209 corrects the deviation for the number of user-generated contents included in the correlation table. The correction value used for the correction is corrected so as to return to zero. The same applies to the case where the width of the distribution is wide. However, when the spread of the distribution cannot be corrected, the correlation table correction unit 209 may use a correction parameter for additional correction. For example, the correlation table correction unit 209 may perform correction to increase the correction parameters by performers and reduce the number of tweets included in the correlation table when a popular performer appears. The correction of the correlation table by the correlation table correction unit 209 based on the difference between the audience rating estimated by the audience rating estimation unit 206 and the audience rating calculated by the external device 4 is calculated by the external device 4. This is realized by measuring the TV program audience rating, the change in the number of tweets, and the correlation rate in the correlation table, and repeatedly performing the deviation width feedback.

  In the present embodiment, the description has been made based on the number of tweets of the short text posting site as the user-generated content, but any other user-generated content can be used. The logic of audience rating estimation is to create a correlation table equivalent to that for short posting sites, but depending on the user-generated content, it is not a simple number such as the number of tweets, but the strength of correlation and reliability. In some cases, the correlation with the audience rating is measured. In this case, the accuracy of audience rating estimation varies depending on each user-generated content. Therefore, the estimation accuracy is measured, and the audience rating estimated for each user-generated content is weighted. For example, when the reliability of the user-generated content A is 50 and the reliability of the user-generated content B is 100, the estimated audience rating is weighted at a ratio of user-generated content A: user-generated content B = 1: 2. A composite audience rating estimate is calculated. Thereby, the estimation accuracy of the audience rating can be improved.

  As a specific configuration when calculating the above-mentioned composite audience rating estimation value, the storage unit 204 is configured for each type of user-generated content related to a television program, specifically, for each bulletin board, blog, video sharing site, SNS, etc. A correlation table is stored in advance. The target content number acquisition unit 205 acquires the number of user-generated content for each type of user-generated content related to the target television program. The audience rating estimation unit 206 includes the number of user-generated contents for each type of user-generated content related to the target television program acquired by the target content number acquisition unit 205 and the user generation related to the TV program stored in advance by the storage unit 204. Based on the correlation table for each content type, the audience rating of the television program for each type of user-generated content related to the television program is estimated. Then, the audience rating estimation unit 206 determines the television program based on the estimated audience rating of the television program for each type of user-generated content related to the television program and the weighting for each type of user-generated content related to the television program. Estimate the audience rating.

  Next, processing of the audience rating estimation method by the audience rating estimation system 1 of the present embodiment will be described with reference to FIGS. FIG. 7 is a sequence diagram showing a process of generating a correlation table among the processes of the audience rating estimation method, and FIG. 8 is a viewer rating based on a correlation table stored in advance among the processes of the audience rating estimation method. It is a sequence diagram which shows the process which estimates this.

  First, the flowchart of FIG. 7 will be described. First, the content number acquisition unit 201 acquires user-generated content (UGC) (step S1). Next, the television program targeted by the acquired user-generated content is specified by the content number acquisition unit 201 (step S2). Next, the number-of-contents acquisition unit 201 acquires the number of user-generated contents for each identified television program (step S3).

  Subsequently, the audience rating is acquired from the external device 4 by the audience rating acquisition unit 202 (step S4). Next, the correlation table generation unit 203 analyzes the characteristics of the user-generated content acquired in step S1 (step S5). Next, the correlation table generation unit 203 analyzes the characteristics of the television program specified in step S2 (step S6). Next, the correlation table generation unit 203 acquires in step S3 based on the result of the characteristic analysis of the user-generated content performed in step S5 and the result of the characteristic analysis of the television program performed in step S6. The number of user-generated contents thus corrected is corrected (step S7). Next, the correlation table generation unit 203 generates a correlation table based on the audience rating acquired in step S4 and the number of user-generated contents corrected in step S7 (step S8). Next, the storage unit 204 stores a correlation table (step S9, storage step).

  Next, the flowchart of FIG. 8 will be described. First, the target content number acquisition unit 205 acquires user-generated content related to a target TV program whose audience rating is estimated (step S20). Next, the number of acquired user-generated contents is acquired by the target content number acquisition unit 205 (step S21, target content number acquisition step). Next, the audience rating estimation unit 206 analyzes the characteristics of the user-generated content acquired in step S20 (step S22). Next, the audience rating estimation unit 206 analyzes the characteristics of the target TV program whose audience rating is to be estimated (step S23). Next, the audience rating estimation unit 206 obtains in step S21 based on the result of the characteristic analysis of the user-generated content performed in step S22 and the result of the characteristic analysis of the television program performed in step S23. The number of user-generated contents thus corrected is corrected (step S24). Next, the audience rating is estimated by the audience rating estimation unit 206 based on the correlation table stored in advance and the number of user-generated contents corrected in step S24 (step S25, audience rating estimation step). .

  Next, processing of the above-described composite audience rating estimation value calculation method will be described with reference to FIG. As a precondition, it is assumed that user-generated content A and user-generated content B are stored in the server 6 as types of user-generated content. First, the audience rating based on the user-generated content A is estimated by the audience rating estimation device 2 (step S40). Next, the audience rating based on the user-generated content B is estimated by the audience rating estimation device 2 (step S41). Note that the audience rating is estimated in step S40 and step S41 according to the processing shown in steps S20 to S25 described above. Next, the audience rating is estimated based on the audience rating estimated in step S40, the audience rating estimated in step S41, the weight of the user-generated content A, and the weight of the user-generated content B by the audience rating estimation unit 206. Is estimated (step S42).

  Next, the effect of the audience rating estimation apparatus 2 configured as described above will be described.

  According to the audience rating estimation device 2 of the present embodiment, a correlation table between the number of user-generated content related to the television program generated on the network 3 by the viewer of the television program and the audience rating of the television program is stored. Stored in advance by the unit 204. Based on the number of user-generated contents related to the target television program acquired by the target content number acquisition unit 205 and the correlation table stored in advance by the storage unit 204, the audience rating estimation unit 206 views the television program. The rate is estimated. By adopting such a configuration, it is possible to quickly estimate the audience rating of the TV program by acquiring the number of user-generated contents related to the TV program being broadcast. For example, since the audience rating is estimated based on the correlation table between the number of past user-generated contents stored in advance and the audience rating, a more accurate audience rating of the television program can be estimated.

  Further, according to the audience rating estimation device 2 of the present embodiment, the storage unit 204 stores in advance a correlation table between the number of user-generated contents for each predetermined time and the audience rating for each predetermined time, and acquires the number of target contents. The unit 205 preferably acquires the number of user-generated contents in a predetermined time. In addition, the audience rating estimation unit 206 is based on the number of user-generated content related to the target television program acquired by the target content number acquisition unit 205 and the correlation table stored in advance by the storage unit 204 based on the number of user-generated content. It is preferable to estimate the audience rating of the television program at a predetermined time. By adopting such a configuration, it is possible to quickly estimate the audience rating of the TV program at the predetermined time by acquiring the number of user-generated contents for the predetermined time regarding the TV program being broadcast. Further, for example, since the audience rating is estimated based on the correlation table between the number of past many user-generated contents for each predetermined time stored in advance and the audience rating for each predetermined time, a more accurate television at the predetermined time is obtained. The audience rating of the program can be estimated.

  Also, according to the audience rating estimation device 2 of the present embodiment, the content number acquisition unit 201 that acquires the number of user-generated content related to the TV program generated on the network by the viewer of the TV program, and the external device calculate The audience rating acquisition unit 202 that acquires the audience rating of the television program that has been received, the number of user-generated content related to the television program acquired by the content number acquisition unit 201, and the audience rating of the television program acquired by the audience rating acquisition unit 202 It is preferable to further include a correlation table generation unit 203 that generates a correlation table with. The storage unit 204 preferably stores the correlation table generated by the correlation table generation unit 203 in advance. With this configuration, for example, a correlation table can be generated from the number of user-generated contents in the past and the audience rating, and the audience rating is estimated based on the correlation table. The audience rating can be estimated.

  Further, according to the audience rating estimation apparatus 2 of the present embodiment, the correlation table generation unit 203 is configured to store the user-generated content acquired by the content number acquisition unit 201 based on the characteristics of the user-generated content or the characteristics of the TV program. It is preferable to correct the number. By adopting such a configuration, it is possible to estimate a more accurate audience rating of a television program in consideration of the characteristics of user-generated content or the characteristics of a television program.

  In addition, according to the audience rating estimation device 2 of the present embodiment, the audience rating estimation unit 206 is configured such that the user-generated content acquired by the target content number acquisition unit 205 based on the characteristics of the user-generated content or the characteristics of the TV program. It is preferable to correct the number. By adopting such a configuration, it is possible to estimate a more accurate audience rating of a television program in consideration of the characteristics of user-generated content or the characteristics of a television program.

  Further, according to the audience rating estimation device 2 of the present embodiment, the storage unit 204 stores in advance based on the difference between the audience rating estimated by the audience rating estimation unit 206 and the audience rating calculated by the external device 4. And a correlation table correction unit 209 that corrects the correlation table. The storage unit 204 preferably stores the correlation table corrected by the correlation table correction unit 209. By adopting such a configuration, it is possible to correct the correlation table with higher accuracy by feeding back the audience rating estimation result, and thereby to estimate the audience rating of the television program more accurately.

  Further, according to the audience rating estimation apparatus 2 of the present embodiment, the storage unit 204 stores a correlation table in advance for each category for classifying television programs, and the target content number acquisition unit 205 performs the target for each category. The audience rating estimation unit 206 obtains the number of user-generated content related to the target TV program for each category acquired by the target content number acquisition unit 205 and the storage unit 204 in advance. Based on the stored correlation table for each category, the audience rating of the television program for each category is estimated, and based on the estimated audience rating of the television program for each category, the audience rating of the television program is estimated. Is preferable. If such a configuration is adopted, the audience rating is estimated based on the audience rating estimated for each category in classifying the television program, so that the more accurate audience rating of the television program can be estimated.

  Also, according to the audience rating estimation device 2 of the present embodiment, the storage unit 204 stores a correlation table in advance for each type of user-generated content related to a television program, and the target content number acquisition unit 205 relates to the target television program. It is preferable to acquire the number of user-generated contents for each type of user-generated content. The audience rating estimation unit 206 also relates to the number of user-generated content for each type of user-generated content related to the target TV program acquired by the target content number acquisition unit 205 and the TV program stored in advance by the storage unit 204. It is preferable to estimate the audience rating of the television program for each type of user-generated content related to the television program based on the correlation table for each type of user-generated content. Then, the audience rating estimation unit 206 determines the television program based on the estimated audience rating of the television program for each type of user-generated content related to the television program and the weighting for each type of user-generated content related to the television program. It is preferable to estimate the audience rating. With this configuration, since the audience rating is estimated based on the audience rating estimated for each type of user-generated content and the weighting for each type of user-generated content, it is possible to estimate the audience rating of a more accurate television program. it can.

  In addition, according to the audience rating estimation device 2 of the present embodiment, the audience rating estimation unit 206 analyzes the contents of the user-generated content used for estimating the audience rating, and the audience rating estimation unit 206 estimates it. It is preferable to further include an output unit 208 that outputs the received audience rating and the analysis result by the analysis unit 207. By adopting such a configuration, the analysis result of the content of the user-generated content can be output together with the audience rating, so that the user can know the viewer's reaction in more detail.

DESCRIPTION OF SYMBOLS 1 ... Audience rating estimation system, 2 ... Audience rating estimation apparatus, 3 ... Network, 4 ... External apparatus, 5 ... Television, 6 ... Server, 7 ... Personal computer, 201 ... Content number acquisition part, 202 ... Audience rating acquisition part, 203 ... correlation table generation unit, 204 ... storage unit, 205 ... target content number acquisition unit, 206 ... audience rating estimation unit, 207 ... analysis unit, 208 ... output unit, 209 ... correlation table correction unit.

Claims (10)

Storage means for storing in advance a correlation table between the number of user-generated contents related to the TV program generated on the network by viewers of the TV program and the audience rating of the TV program;
Target content number acquisition means for acquiring the number of user-generated content related to the target TV program;
Audience rating estimation means for estimating the audience rating of the TV program based on the number of user-generated contents related to the target TV program acquired by the target content count acquisition means and the correlation table stored in advance by the storage means. When,
An audience rating estimation apparatus. The storage means stores in advance a correlation table between the number of user-generated contents every predetermined time and the audience rating every predetermined time,
The target content number acquisition means acquires the number of user-generated content in a predetermined time,
The audience rating estimation means is configured to determine the predetermined time based on the number of user-generated contents regarding the target television program acquired by the target content number acquisition means and the correlation table stored in advance by the storage means. Estimate the audience rating of the TV program at
The audience rating estimation apparatus according to claim 1. Content number acquisition means for acquiring the number of user-generated content related to the TV program generated on the network by the viewer of the TV program;
Audience rating acquisition means for acquiring the audience rating of a TV program calculated by an external device;
Correlation table generating means for generating a correlation table between the number of user-generated contents related to the television program acquired by the content number acquiring means and the audience rating of the television program acquired by the audience rating acquiring means;
Further comprising
The storage means stores in advance the correlation table generated by the correlation table generation means;
The audience rating estimation apparatus according to claim 1 or 2, wherein The correlation table generation means corrects the number of user-generated contents acquired by the content number acquisition means based on the characteristics of user-generated contents or the characteristics of a television program.
The audience rating estimation apparatus according to claim 3. The audience rating estimation means corrects the number of user-generated contents acquired by the target content number acquisition means based on the characteristics of user-generated contents or the characteristics of television programs.
The audience rating estimation apparatus according to any one of claims 1 to 4, wherein Correlation table correction means for correcting the correlation table stored in advance by the storage means based on the difference between the audience rating estimated by the audience rating estimation means and the audience rating calculated by the external device;
Further comprising
The storage means stores the correlation table corrected by the correlation table correction means;
The audience rating estimation apparatus according to any one of claims 1 to 5, wherein The storage means stores in advance a correlation table for each category in classifying television programs,
The target content number acquisition means acquires the number of user-generated content related to the target TV program for each of the categories,
The audience rating estimation means is based on the number of user-generated contents related to the target TV program for each category acquired by the target content number acquisition means and the correlation table for each category stored in advance by the storage means. Estimating the audience rating of the television program for each category, and estimating the audience rating of the television program based on the estimated audience rating of the television program for each category.
The audience rating estimation apparatus according to any one of claims 1 to 6, wherein The storage means stores a correlation table in advance for each type of user-generated content related to a television program,
The target content number acquisition means acquires the number of user-generated content for each type of user-generated content related to the target TV program,
The audience rating estimation means includes the number of user-generated contents for each type of user-generated content related to the target television program acquired by the target content number acquisition means, and the user generation related to the TV program stored in advance by the storage means. Based on the correlation table for each type of content, the audience rating of the TV program for each type of user-generated content regarding the TV program is estimated, and the TV program for each type of user-generated content regarding the estimated TV program is estimated. Estimating the audience rating of the TV program based on the audience rating and the weighting of each type of user-generated content related to the TV program;
The audience rating estimation apparatus according to any one of claims 1 to 7, wherein Analyzing means for analyzing the content of user-generated content used in estimating the audience rating by the audience rating estimating means;
Output means for outputting the audience rating estimated by the audience rating estimating means and the analysis result by the analyzing means;
The audience rating estimation apparatus according to any one of claims 1 to 8, further comprising: An audience rating estimation method by an audience rating estimation device,
A storage step in which the storage means of the audience rating estimation device stores in advance a correlation table between the number of user-generated content related to the television program generated on the network by the viewer of the television program and the audience rating of the television program; ,
A target content number acquiring means for acquiring the number of user-generated content related to the target television program, wherein the target content number acquiring means of the audience rating estimation device;
The audience rating estimation means of the audience rating estimation device, based on the number of user-generated contents related to the target television program acquired in the target content count acquisition step and the correlation table stored in advance in the storage step, An audience rating estimation step for estimating an audience rating of a TV program;
Audience rating estimation method.

Images