Journal of Digital Information Management L. Lakshmi, P. Bhaskara Reddy, C.Shoba Bindu 15 3 2017 http://dline.info/fpaper/jdim/v15i3/jdimv15i3_2.pdf Queries submitted by users to search engines might be ambiguous, concise and their meaning may change over time. Web query classification is emphasized by various search engines nowadays due to the increase in the size of the web, as millions of web pages are added to it every day. Some of the current Information Retrieval (IR) systems like Library Online Public Access Catalog (OPAC), dialog system and numerous web search engines need classical Boolean approaches in addition to the current supervised methods. Document retrieval is a process which mainly involves retrieval of relevant documents for user queries and matching the results using efficient algorithms like 'page rank' and 'learn to rank' algorithms. In this paper, to retrieve more relevant documents against user query by reducing non-relevant documents, we proposed a tree based classification of web queries using Simple Ordinal Classification (SOC) and navigation of search keywords is performed dynamically with search session time of users. This method reduces the retrieval of most of the non-relevant documents and navigation cost using efficient B-tree data structure. It provides documents that match all keywords present in the user query and best resulting web pages for users with different categories of interest. We built a prototype application to evaluate the proposed approach. Our experimental results revealed that SOC has significant performance improvement with existing approaches.