Information processing apparatus, method for processing information, and program

Abstract

An information processing apparatus includes a specific-evaluation information acquisition unit that acquires an evaluation of a predetermined content item as a specific evaluation, the evaluation having been input by a user in accordance with an ordinal scale; a language-evaluation information extraction unit that acquires a language evaluation from language information regarding an evaluation sentence in which an evaluation of the predetermined content item is expressed in a language, the evaluation sentence having been input by the user; and a recommendation unit that recommends a content item that matches the user's preference in accordance with whether the specific evaluation is a positive or negative evaluation and whether the language evaluation is a positive or negative evaluation.

1 . An information processing apparatus, comprising:
 specific-evaluation information acquisition means for acquiring an evaluation of a predetermined content item as a specific evaluation, the evaluation having been input by a user in accordance with an ordinal scale;   language-evaluation information extraction means for acquiring a language evaluation from language information regarding an evaluation sentence in which an evaluation of the predetermined content item is expressed in a language, the evaluation sentence having been input by the user; and   recommendation means for recommending a content item that matches the user's preference in accordance with whether the specific evaluation is a positive or negative evaluation and whether the language evaluation is a positive or negative evaluation.   
     
     
         2 . The information processing apparatus according to  claim 1 , further comprising:
 evaluation correction means for correcting a main evaluation, which is the specific evaluation or the language evaluation, in accordance with whether the specific evaluation is a positive or negative evaluation and whether the language evaluation is a positive or negative evaluation; and   preference learning means for learning the user's preference by using the main evaluation corrected by the evaluation correction means,   wherein the recommendation means recommends a content item that matches the user's preference in accordance with preference information representing the user's preference, the preference information being created by learning performed by the preference learning means.   
     
     
         3 . The information processing apparatus according to  claim 2 ,
 wherein the evaluation correction means corrects the main evaluation in such a manner that the main evaluation is given a heavier weight in a case in which both the specific evaluation and the language evaluation are positive evaluations or negative evaluations, and the evaluation correction means corrects the main evaluation in such a manner that the main evaluation is given a lighter weight in a case in which one of the specific evaluation and the language evaluation is a positive evaluation and the other is a negative evaluation.   
     
     
         4 . The information processing apparatus according to  claim 3 ,
 wherein the language-evaluation information extraction means also acquires a level of the language evaluation from the language information regarding the evaluation sentence, and   wherein the evaluation correction means corrects the main evaluation in such a manner that the main evaluation is given a heavier weight in accordance with a degree corresponding to the level of the language evaluation in a case in which both the specific evaluation and the language evaluation are positive evaluations or both of them are negative evaluations, and the evaluation correction means corrects the main evaluation in such a manner that the main evaluation is given a lighter weight in accordance with the degree corresponding to the level of the language evaluation in a case in which one of the specific evaluation and the language evaluation is a positive evaluation and the other is a negative evaluation.   
     
     
         5 . The information processing apparatus according to  claim 2 , further comprising:
 determination means for determining which of the specific evaluation and the language evaluation is to be the main evaluation,   wherein the preference learning means learns the user's preference by using the specific evaluation and also learns the user's preference by using the language evaluation, and   wherein the determination means determines the main evaluation in accordance with a prediction error between a predicted evaluation, which is obtained in accordance with the preference information created by learning using the specific evaluation, of the predetermined content item and the specific evaluation and in accordance with a prediction error between a predicted evaluation, which is obtained in accordance with the preference information created by learning using the language evaluation, of the predetermined content item and the language evaluation.   
     
     
         6 . The information processing apparatus according to  claim 1 ,
 wherein the language-evaluation information extraction means also acquires an attribute of the language evaluation from the language information regarding the evaluation sentence, and   wherein the recommendation means determines a content feature value, which is to be used to calculate a degree of similarity or a distance between the predetermined content item and a recommendation-candidate content item, in accordance with a content feature value corresponding to the attribute and in accordance with whether the specific evaluation is a positive or negative evaluation and whether the language evaluation is a positive or negative evaluation, and recommends the recommendation-candidate content item as the content item that matches the user's preference, in accordance with the degree of similarity or the distance between the predetermined content item and the recommendation-candidate content item when the determined content feature value is used.   
     
     
         7 . The information processing apparatus according to  claim 6 ,
 wherein the recommendation means determines a feature value other than the content feature value corresponding to the attribute to be the content feature value, which is to be used to calculate the degree of similarity or the distance, in a case in which the specific evaluation is a positive evaluation and the language evaluation is a negative evaluation, and the recommendation means determines the content feature value corresponding to the attribute to be the content feature value, which is to be used to calculate the degree of similarity or the distance, in a case in which the specific evaluation is a negative evaluation and the language evaluation is a positive evaluation.   
     
     
         8 . The information processing apparatus according to  claim 1 , further comprising:
 preference learning means for learning the user's preference by using the language evaluation or the specific evaluation, and a feature value of the predetermined content item,   wherein the language-evaluation information extraction means also acquires an attribute of the language evaluation from the language information regarding the evaluation sentence, and   wherein the preference learning means determines a learning method in accordance with a content feature value corresponding to the attribute and in accordance with whether the specific evaluation is a positive or negative evaluation and whether the language evaluation is a positive or negative evaluation, and learns the user's preference by the learning method.   
     
     
         9 . The information processing apparatus according to  claim 8 ,
 wherein the preference learning means determines that the learning method is to be a method in which the user's preference is learnt by weighting a feature value of the predetermined content item, the feature value corresponding to the attribute, in a case in which the specific evaluation and the language evaluation are positive evaluations, determines that the learning method is to be a method in which the user's preference is learnt by using only a feature value of the predetermined content item, the feature value being other than the feature value corresponding to the attribute, in a case in which the specific evaluation is a positive evaluation and the language evaluation is a negative evaluation, and determines that the learning method is to be a method in which the user's preference is learnt by using only the feature value of the predetermined content item, the feature value corresponding to the attribute, in a case in which the specific evaluation is a negative evaluation and the language evaluation is a positive evaluation.   
     
     
         10 . A method for processing information performed by an information processing apparatus, the method comprising the steps of:
 acquiring an evaluation of a predetermined content item as a specific evaluation, the evaluation having been input by a user in accordance with an ordinal scale;   acquiring a language evaluation from language information regarding an evaluation sentence in which an evaluation of the predetermined content item is expressed in a language, the evaluation sentence having been input by the user; and   recommending a content item that matches the user's preference in accordance with whether the specific evaluation is a positive or negative evaluation and whether the language evaluation is a positive or negative evaluation.   
     
     
         11 . A program for causing a computer to execute processing including the steps of:
 acquiring an evaluation of a predetermined content item as a specific evaluation, the evaluation having been input by a user in accordance with an ordinal scale;   acquiring a language evaluation from language information regarding an evaluation sentence in which an evaluation of the predetermined content item is expressed in a language, the evaluation sentence having been input by the user; and   recommending a content item that matches the user's preference in accordance with whether the specific evaluation is a positive or negative evaluation and whether the language evaluation is a positive or negative evaluation.   
     
     
         12 . An information processing apparatus, comprising:
 a specific-evaluation information acquisition unit that acquires an evaluation of a predetermined content item as a specific evaluation, the evaluation having been input by a user in accordance with an ordinal scale;   a language-evaluation information extraction unit that acquires a language evaluation from language information regarding an evaluation sentence in which an evaluation of the predetermined content item is expressed in a language, the evaluation sentence having been input by the user; and   a recommendation unit that recommends a content item that matches the user's preference in accordance with whether the specific evaluation is a positive or negative evaluation and whether the language evaluation is a positive or negative evaluation.