CPC G06F 16/3334 (2019.01) [G06F 16/3347 (2019.01); G06Q 50/01 (2013.01)] | 19 Claims |
1. A method for analyzing sentiment, the method comprising:
identifying a plurality of comments;
providing, to a topic model, a first comment of the plurality of comments and a value k indicative of k topics to be identified via the topic model, wherein k is an integer greater than 1;
generating, by the topic model, a topic vector for the first comment, the topic vector including k elements, each of the k elements indicative of a degree in which the first comment is associated with a corresponding topic of the k topics;
identifying a first element of the k elements of the topic vector associated with a first topic of the k topics;
determining whether the first element exceeds a weight threshold;
in response to the determining that the first element exceeds the weight threshold, classifying the first comment into the first topic;
estimating a first sentiment score for the first comment;
in response to classifying the first comment into the first topic, and in response to estimating the first sentiment score for the first comment, associating the first sentiment score to the first topic and to a temporal value;
generating a time series of sentiment values for the first topic including the first sentiment score; and
triggering an action in response to generating the time series of sentiment values.
|