意见挖掘中特征权重的计算方法研究

Research on the Calculation Method of the Features' Weights in the Feature-based Opinion Mining

基于特征的意见挖掘中,特征权重计算是一个难点。在对各种权重计算方法深入研究的基础上,提出一种新的权重计算方法以解决该问题。该方法集成了层次分析法和G1法的基本思想,可以有效提高权重计算的准确性。一方面,该方法使用G1法对层次分析法中单层次指标权重计算进行改进,以克服层次分析法中判断矩阵的不一致性带来的问题;另一方面,在单层次指标的层次总权重计算中,提出一个总权重计算通用公式,扩展了层次分析法中层次总权重计算公式,使其更适合一般的应用。

In the feature based Opinion mining, the calculation of features＂ weight is a difficult. This paper, based on the deep study of the weight calculation methods, proposes a new method to calculate the weight of features, which mainly integrates the thought of the fuzzy analytic hierarchy process and the G1 method. This method can effectively improve the accuracy of the weight, that is because, the new method has improved the AHP from two aspects： first, it uses the G1 method to overcome the inconsistency of the judgment matrix; second, in order to making the AHP more suitable for general use, it uses a new general formula to calculate the whole weight of features instead of the method used in the AHP.