Effective Estimation of Context Similarity: A Proposed Matching Model Based on Weighted Semantic Load

In this paper, we propose a new model to calculate the similarity of two sentences. The proposed scheme is based on the amount of semantic load which is shared between two sentences. Since verb is the essential part of a sentence, the main focus of the proposed model is on the verbs of two sentences. We supposed the verb as the anchor of the sentence which carries the most semantic of the sentence. The proposed modeldepends on part of speech (POS), the partial order of words in the sentence and the words’ senses. Theresults by Precision and Recall are promising and benchmarks show that the proposed method improves the quality of the retrieved matched sentences.