Query Interface Classification Based on Singular Value Decomposition

• Published in: 2015 8th International Symposium on Computational Intelligence and Design (ISCID) Vol. 1; pp. 572 - 575
• Main Authors: Heping Gou, Yongxia Jing, Yuan Zhao
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2015
• Subjects: classification; Feature extraction; Matrix decomposition; query interface; Semantics; Singular value decomposition; Text categorization; Training
• DOI: 10.1109/ISCID.2015.198
• ISBN: 9781467395861; 1467395862

Deep Web query interface classification is the main method to realize classification of Deep Web database, but because of the heterogeneity and diversity of information systems, the query interfaces on Deep Web online database in same domain are diversiform, which leads to the increased number of synonyms and high dimensional feature vectors of the query interface. In this paper, the query interface classification method based on singular value decomposition is proposed to realize semantic translation of query interface feature and data dimension reduction. It converts the query interface on original semantic space to another semantic space with a low rank matrix to approximate the original matrix. The experiments show that the method can reduce the classification overhead, improve the efficiency and accuracy of Deep Web online database retrieval.