APPROVAL RATING PREDICTION SYSTEM USING NEWS DATA

摘要

The present invention relates to a system for predicting approval ratings using news data, comprising: a data collection module for collecting a plurality of news data related to a keyword representing a specific person; A preprocessing module for generating first data by removing words not related to the keyword from the news data; A news extraction module for generating second data by extracting news data related to an approval rating of a specific person from the first data using a topic modeling method; And an approval rate calculation module that analyzes the emotion information of the second data and calculates approval rate information, which is information on the approval rate of a specific person, wherein the emotion information is based on an emotion index expressing the reader’s emotion for the news as an emoticon Information, and the approval rate calculation module calculates an NSI value, which is a numerical value by analyzing the emotion information of the second data, and classifies the second data according to the NSI value to generate third data ( News Sentiment Index) calculation module, DAR (Daily Approval Rating) calculation module for calculating the approval rate for each date based on the third data, and AR (Approval Rating) calculation module for calculating the approval rate information based on the approval rate for each date. Includes. According to the present invention, it is possible to grasp the approval rate of the public for a specific person in real time, and it is possible to quickly grasp the approval rate for a specific person at low cost.