Semantic Similarity Computation Based on Multi-feature Combination using HowNet

Abstract

Semantic similarity between words is becoming a generic problems for many applications of computational linguistics and artificial intelligence. The difficulty lies in how to develop a computational method that is capable of generating satisfactory results close to how humans perceive. This paper proposes a semantic similarity approach that is based on multi-feature combination. One of the benchmarks is Miller and Charles’ list of 30 noun pairs which had been manually designated similarity measurements. We correlate our experiments with those computed by several other methods. Experiments on Chinese word pairs show that our approach is close to human similarity judgments.