본 논문에서는 이중 주어를 가진 문장에서 오피니언 마이닝의 중요 요소인 감정의 평가 요소와 감정 특징간의 결합을 수행할 수 있는 방법을 제안한다. 오피니언 마이닝의 감정 평가 요소...
본 논문에서는 이중 주어를 가진 문장에서 오피니언 마이닝의 중요 요소인 감정의 평가 요소와 감정 특징간의 결합을 수행할 수 있는 방법을 제안한다.
This thesis presents how to combine Sentiment Entity with Sentiment Features, which are key factors in Opinion Mining. In Opinion Mining, Sentiment Entity, which must be decorated by Sentiment Features found in predicates, is carried out on each Domai... This thesis presents how to combine Sentiment Entity with Sentiment Features, which are key factors in Opinion Mining. In Opinion Mining, Sentiment Entity, which must be decorated by Sentiment Features found in predicates, is carried out on each Domain. If there is no consideration for thispart, problems that Sentiment Features can be uncertain will arise. In the case of a simple sentence in Opinion Mining, these problems cannot have an influence on classification efficiency. However, it is important to consider this part in sentences that cause dual subjects phenomenon. Since thereis no consideration of this part in existing Polarity Sentiment Classification, the context cannot be reflected. Accordingly, there are solutions conceived by grammatical features in Korean to the problems in this thesis. First of all, sentences generating dual subjects phenomenon are separated by a conjunction, the ending of a word and a postposition, which are the principal factors determining grammatical features in Korean. Also, Sentiment Entity, which is subject in input sentences, and Sentiment Features, which is a predicate, are paired off. Nouns that can be Sentiment Features among phrases are chosen by frequency of appearance and how their hierarchy are composed. Also, it determine the ranking. Through these ways, there are suggestions about how to solve refraction in View Point of Opinion mining. ，韩语论文范文，韩语论文 |
