AN INFORMATION-RETRIEVAL METHOD BASED ON N-LEVEL VECTOR MODEL

A new information-retrieval algorithm based on the N-level vector model is proposed. The N-level vector model partitions a document into N level text paragraphs. The text feature vectors and the text weight vectors are defined according to the text paragraphs’ context. The calculation method of the feature vectors and the similarity are defined much more precisely such that the algorithm can adapt the dynamitic extension of the document set. The theoretic analysis and the experimental results show that the new algorithm has higher precision and faster computation speed.