Abstract:
PROBLEM TO BE SOLVED: To further accurately estimate a user's emotion without increase in the user's burden. SOLUTION: A learning part 103 learns, when stimulation information is presented during learning processing, an emotion determination model showing the relation between a characteristic vector showing the user's state supplied from a sensor part and the user's interest based on user instruction information intentionally inputted by the user, the characteristic vector, and an emotion map mutually relating the instruction content and interest of the user. After the learning processing, an estimation part 104 estimates the user's emotion to the stimulation information based on the characteristic vector and the emotion determination model. This invention is applicable to a recommendation system for recommending contents to users. COPYRIGHT: (C)2007,JPO&INPIT
Abstract:
PROBLEM TO BE SOLVED: To use a smaller amount of calculation to retrieve contents, which conform to user taste information, or contents, which are similar to designated contents and to present the retrieved contents to the user. SOLUTION: The items (titles, artist names, categories, review texts, tempos, beats, rhythms and so on) of music metadata are classified into the respective ones of cluster layers (first to n-th layers) and the actual information of the items are classified (clustered) into the respective ones of clusters provided in the cluster layers. Then, cluster information, which comprises cluster IDs (such as CL11 in Fig.2) of the clusters into which the actual information of the items of the metadata are classified, as information indicative of the features of the music, instead of the metadata, is complied into a database. The present invention can be applied, for example, to a server of websites, in which the purchases of contents such as music data are recommended, and to video recorders in which timed recordings of television programs are recommended. COPYRIGHT: (C)2007,JPO&INPIT
Abstract:
PROBLEM TO BE SOLVED: To accurately acquire information interesting to a user and present the information to the user. SOLUTION: Operation history data on a webpage viewed by a user are analyzed and words included in text data extracted from the webpage corresponding to the operation history are analyzed to generate user characteristic information 209, and content is recommended to the user with the user characteristic information. The user characteristic information 209 reflects the history of operations on the webpage interesting to the user, so that content interesting to the user can be recommended. The processing is automatically executed, so that without complicated operations and even without special knowledge of a network and the like, the user can be recommended interesting content. COPYRIGHT: (C)2005,JPO&NCIPI
Abstract:
PROBLEM TO BE SOLVED: To extract data of interest from a large number of apparatuses, and to recommend a program near to an interest of a user. SOLUTION: Preference information of the user is acquired from a user information acquiring part 91 and a history information acquiring part 92, and the data of interest is generated in an interest extraction part 93. Related information is retrieved by a related information retrieving part 94, based on the data of interest, and an evaluation input part 96 receives an evaluation input of the exhibited related information to reflect a result onto the interest extracting part. The data of interest is confirmed of the propriety of transmission by a transmission confirming part 101, and is transmitted by a change operation part 102 to be transmitted to other equipment. The present invention is applicable for a personal computer, an HDD recorder or the like. COPYRIGHT: (C)2005,JPO&NCIPI
Abstract:
PROBLEM TO BE SOLVED: To make it possible to recommend a content to a user, even in a metadata imparted with no classification. SOLUTION: A metadata analysis part 103 resolves the metadata acquired by a metadata acquiring part 101 into respective components, and a dictionary preparing part 104 generates genres, keywords, and dictionary data correlated with the respective components. A related information database preparing part 102 imparts the genres to the metadata not imparted with the genres, referring to the dictionary data, to generate a related information database of the content. A related information database retrieval part 107 specifies the genre based on the keyword of a data of interest, referring to the dictionary data, and retrieves related information to recommend the content to the user. The present invention is applicable for a personal computer or an HDD recorder. COPYRIGHT: (C)2005,JPO&NCIPI
Abstract:
PROBLEM TO BE SOLVED: To generate a positive history vector and a negative history vector. SOLUTION: In a step S71, the negative history vector is generated from initial registration contents, in a step S72, it is determined that the initial registration contents are not changed. When it is determined in a step S73 that an operation log is supplied, whether the supplied operation log is the positive history or not is determined in a step S74. When the supplied operation log is the positive log, in a step S75, a program vector corresponding to the operation log of the positive history is additionally stored as the positive history, and in a step S76, the positive history vector is generated from the sum of the program vectors of the positive history. When the supplied operation log is the negative history, in a step S77, the program vector corresponding to the operation log of the negative history is additionally stored as the negative history, and in a step S78, the negative history vector is generated from the sum of the program vectors of the negative history. The present invention is applied to a program recommendation processor. COPYRIGHT: (C)2004,JPO&NCIPI
Abstract:
PROBLEM TO BE SOLVED: To provide an information processor and an information processing method by which weighting information stipulating the degree of contribution to a plurality of respective items is associated with attribute information on contents in calculating similarity between the attribute information and information on a user's preference, and to provide an information processing system, a recording medium, and a program. SOLUTION: In a step S11, the supply of EPG (Electric Program Guide) data is received; in a step S12, metadata required for generating a program vector PP (Program-to-Program) is extracted; in a step S13, morphological analysis is applied to a title, contents and the like contained in the metadata and the title, the contents and the like are decomposed into words; in a step S14, vectorization is applied to the respective items contained in the metadata to generate the program vector PP; in a step S15, a corresponding effect vector is extracted based on the genre of the program in the supplied metadata; in a step S16, the extracted effect vector is related to a generated vector, and processing is completed. The present invention is applied to a delivery server delivering the contents. COPYRIGHT: (C)2004,JPO&NCIPI
Abstract:
PROBLEM TO BE SOLVED: To generate user's taste information on the basis of attribute information of contents meeting prescribed conditions. SOLUTION: Program vectors of programs being user model generation objects are acquired in a step S231, and a topic inputted by the user is set as a filtering condition in a step S232. Program vectors are filtered on the basis of the filtering condition to extract program vectors meeting the topic in a step S233, and a user model vector is generated from the sum total of extracted program vectors in a step S234. Cosine distances between the user model vector and program vectors of candidate programs for obtaining recommended information are calculated in a step S235, and the prescribed number of programs out of candidate programs of which the program vectors have higher degrees of similarity to the user model vector are extracted in order from the highest degree as recommended information in a step S236. This invention is applicable to program recommendation apparatus . COPYRIGHT: (C)2004,JPO&NCIPI
Abstract:
PROBLEM TO BE SOLVED: To represent relative information at a good timing in accordance with a situation. SOLUTION: In a step S42, the similarity is calculated on the basis of an inner product of a feature vector of a document corresponding to the generation of event, and a feature vector of each topic. When it is determined that the similarity more than a predetermined specific threshold is calculated in the step S43, the topic having the similarity more than the calculated specific threshold, is selected in a step S44. Thereby, the topic can be extracted when the similarly (more than the specific threshold) which is regarded to be similar to the document corresponding to the generation of event, is calculated, without calculating the similarity to all of the topics, and the relative information corresponding to the topic can be represented. COPYRIGHT: (C)2003,JPO
Abstract:
PROBLEM TO BE SOLVED: To efficiently use a network line. SOLUTION: When preparing a data base composed of data acquired through a network, the retrieving processing to be performed for acquiring these data is intermittently performed. When a prescribed application is in an active state and there is an entry to that application, for example, retrieving processing is interrupted and while using the network line released by that interruption, a generated action is processed.