PARAMETERS FOR EVOLUTION AN INFORMATION RETRIVAL SYSTEM

This document explains the key parameters used to evaluate an Information Retrieval System (IRS). It defines precision, recall, response time, relevance, user satisfaction, coverage, ranking quality, and cost-efficiency as major measures of system performance. It also describes how recall and precision are calculated, supported by formulas and examples, and differentiates fallout from recall by focusing on non-relevant retrieval. Additional evaluation metrics such as F-score, MAP, R-precision, Precision@k, and DCG are included. The document further highlights limitations of recall and precision and lists major information scientists who contributed to IRS evaluation, including Cleverdon, Salton, and Saracevic.