A dynamic adjustment algorithm research of sentiment word weight based on context

The emotion tendency of sentiment word is divided into two types: static emotion tendency and dynamic emotion tendency. Basic semantic lexicon is static emotion tendency, in the real context, but it is different between static emotion tendency and dynamic emotion tendency. The paper proposes a method based on degree lexicon, negative lexicon and dependence relationship of sentence elements. The experimental results show dynamic emotion tendency even more reflects text emotion tendency and get best performance in text sentiment classification.