Information processing apparatus, information processing method, and program
AssigneeNs Solutions Corporation
InventorsIsao Sonobe, Takashi Mitsubuchi, Hideaki Tanaka, Hiroaki Takasu, Kazuhiro Yamada, Yasuhiro Mitsuno
Life of this patent
Estimating the term…
Abstract
A plurality of feature values are extracted from input data as document data, distributed representations of words that correspond to the respective extracted plurality of feature values is obtained, and the extracted plurality of feature values are aggregated into a plurality of classifications based on the obtained distributed representation.
Claims16
An information processing apparatus comprising:
- an extractor configured to extract a plurality of feature values from input data as document data;
- an obtainer configured to obtain distributed representations of words that correspond to the respective plurality of feature values extracted by the extractor, wherein the distributed representations are represented by vectors of a plurality of dimensions;
- an aggregator configured to aggregate the plurality of feature values extracted by the extractor into a plurality of classifications based on distances between the vectors of the distributed representations;
- an acceptor configured to accept a specification of a classification including a feature value used for determining an evaluation criterion for document data among the plurality of classifications aggregated by the aggregator;
- and
- a determiner configured to determine the evaluation criterion based on a feature value included in a classification indicated by the specification accepted by the acceptor.
An information processing method executed by an information processing apparatus, the method comprising:
- an extracting step of extracting a plurality of feature values from input data as document data;
- an obtaining step of obtaining distributed representations of words that correspond to the respective plurality of feature values extracted at the extracting step, wherein the distributed representations are represented by vectors of a plurality of dimensions;
- an aggregating step of aggregating the plurality of feature values extracted at the extracting step into a plurality of classifications based on distances between the vectors of the distributed representations;
- an accepting step of accepting a specification of a classification including a feature value used for determining an evaluation criterion for document data among the plurality of classifications aggregated at the aggregating step;
- and
- a determining step of determining the evaluation criterion based on a feature value included in a classification indicated by the specification accepted at the accepting step.
A computer-readable recording medium that records a program for causing a computer to execute:
- an extracting step of extracting a plurality of feature values from input data as document data;
- an obtaining step of obtaining distributed representations of words that correspond to the respective plurality of feature values extracted at the extracting step, wherein the distributed representations are represented by vectors of a plurality of dimensions;
- an aggregating step of aggregating the plurality of feature values extracted at the extracting step into a plurality of classifications based on distances between the vectors of the distributed representations;
- an accepting step of accepting a specification of a classification including a feature value used for determining an evaluation criterion for document data among the plurality of classifications aggregated at the aggregating step;
- and
- a determining step of determining the evaluation criterion based on a feature value included in a classification indicated by the specification accepted at the accepting step.
Dependencies are inferred from the claim text.