The research on analyzing emotional tendency of sentence based on watershed algorithm

摘要

The study object of the analysis of sentence emotional tendency is sentences appearing in a particular context, its task is analyzing and extracting a variety of subjective information of the sentences and judging the emotional tendency of the sentences. Existing methods of analysis usually tend to divide the emotional tendency of sentences into positive and negative emotion categories, not further given the extent of positive and negative. This paper presents a method for calculating the emotional tendency of sentences; it builds the different lexicon according to the emotional words which have different parts of speech and different lexicon have different emotional value. According to the emotional score of sentences, the threshold of different emotional tendency can be determined with the help of watershed algorithm. Further, we can use this threshold to determine the emotional tendency of the sentences. The experimental results show that the approach is qualitatively quite useful when used to analyze the sentence emotional tendency in the actual WEB public sentiment system.