This research suggests another ways to recommend movie by analyzing preference inclination depending on usage motivation of users with little profile information, using movie reviews. For the research, the analysis of affective words in movie reviews ...
This research suggests another ways to recommend movie by analyzing preference inclination depending on usage motivation of users with little profile information, using movie reviews. For the research, the analysis of affective words in movie reviews and the analysis of the situation that shows watching movies are conducted in this thesis. First of all, for procuring affective words, totally 68 affective words are chosen by professional consultations and surveys from 834 Korean affective words. Also, for investigating the degree of similarity or dissimilarity among several affective words, the correlation among several affective words is analyzed using multi-dimensional scaling. Second, by conducting surveys about the most accurate situation for watching applied movie, this research collects recommendation situation classified by applied movie. Also, by extracting each affective word from the review of applied movie, this research consequentially finds grounds that show what kind of affective word in movie, which makes users feel satisfied, is worth to be recommended depending on users` situation. Therefore, this research is expected to be a great help for movie recommendation system that is going to be produced by showing the procedure that suggests the dispersion of affective words from the movies which is worth to be recommended depending on users` situation.