شماره ركورد كنفرانس :
3752
عنوان مقاله :
New term weighting schema for improving precision in textual information retrieval Introduction of augmented weight for better retrieval efficiency
پديدآورندگان :
Akbari Danial danial_akbari33@hotmail.com Faculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran , Rastegari Hamid rastegari@iaun.ac.ir Faculty of Computer Engineering, Najafabad Branch, Islamic Azad University, Najafabad, Iran
كليدواژه :
Textual information retrieva l. Random term weights . Heuristic term , weighting scheme
عنوان كنفرانس :
اولين كنفرانس بين المللي مهندسي و علوم كامپيوتر
چكيده فارسي :
Term weighting schema plays a vital role in retrieval efficiency. Evolutionary weighting schemas need full judged document collections for training a proper weighting schema while most of test or real collections are not fully judged. Additionally, a weighting schema that doesn’t learn from previous results will be obsolete in the future. Proposed weighting schema uses famous TF-IDF term weighting schema as default weighting schema, measures augmented weight for each term by analyzing training query set, and manipulate default weights according to augmented weights. This schema improves mean average precision by 1.521372, 4.96126, 6.73124 percent, for top 10, 20 and 30 retrieved document compared to TF-IDF, under CISI document collection.
