Similarity Measure Based on Improved Optimal Assignment Model

Measuring similarity has a wide range of application in information retrieval, machine translation or other related fields. In this paper, we proposed a text similarity computation based on improved optimal assignment model, which combine the improved Hungarian algorithm with the semantic similarity of terms to obtain the maximum semantic similarity between two documents or between a query and a document. Experiment shows that the algorithm has a significant improvement for semantic similarity comparing to traditional models of similarity measure. the method can be applied to document clustering, which will enchance the accuracy of result.