Study on Similar Engineering Decision Problem Identification Based On Combination of Improved Edit-Distance and Skeletal Dependency Tree with POS

To improve the efficiency of solution analyzing of a new engineering decision problem in decision support system (DSS), this paper utilizes sentence similarity computing to identify solved engineering decision problems in DSS which new problems users input are similar to. On the basis of existing methods on Chinese sentence similarity computing and analyzing the characteristics of engineering decision problems expressed in natural language form, an approach with part of speech (POS) for similar engineering decision problem identification is proposed combining the methods based on improved edit-distance and skeletal dependency tree. According to the approach, the sentence similarity is defined as a weighted combination of expression similarity and semantic similarity. Finally, an instance is presented to illuminate the approach and experiment results are shown to evaluate the accuracy of the approach. Compared with other methods, the approach gives a better performance in identifying similar engineering decision problems.