Construction and quantization for a basic sentiment lexicon

In recent years, along with the rapid development of network and universal uses for computers, more and more people choose to share their emotions, feelings and ideas by online reviews. Therefore, some researches for opinion mining and sentiment analysis have become more important. Of course, as a premise for sentiment analysis, it´s necessary for us to establish a basic sentiment lexicon. Because a good sentiment lexicon would plays a critical role in sentiment analysis. Based on predecessors´ thoughts, this paper does some improvements and proposes a novel quantized approach. In addition, a new method is also proposed to improve the accuracy rate for polarity judgment. Experimental results show that our approach does more effective than previous approaches.