This work discusses need and methods of Web search evaluation. The work covers an excellent review of different efforts made in the area. The work discusses the use of different classical content based techniques like Vector Space Model, Boolean Similarity Measures and connectivity based techniques like PageRank for Web search evaluation. The work emphasizes the importance of user feedback based evaluation. But, at the same time, it points out the limitation of user feedback based evaluation in terms of scalability and cost. The work discusses a comprehensive Web search evaluation system where different evaluation techniques are aggregated using rank aggregation techniques. The work discusses in detail many rank aggregation methods for the Web. Finally, the work discusses the architecture of an automatic Web search evaluation system. In this system, different content and connectivity based techniques are combined using rough set based rank aggregation. In rough set based rank aggregation, ranking rules are learnt from the user feedback for the queries in the training set using rough set theory. These rules are then used for combining different Web search evaluation techniques.