The recent explosion in the usage of web services and information technologies had always led to the hot issue of interoperability among heterogeneous systems. Intelligent resources organization in such systems is closely coupled to the principles of Semantic (Intelligent) Web. One of the existing widespread applications of semantic web is Electronic learning (E-learning) service which is deemed to be “Education at all stages”. The prerequisite of this semantic-driven resource management and content delivery in E-learning service has been facilitated by building Ontology thus playing a key role in managing and distributing the resources semantically. Out of the several challenging issues present in the construction of ontology, the problem of anaphora resolution (resolving pronouns) has a greater impact on the expressiveness of the constructed ontology. Anaphora resolution in the web documents with highly related sentences has obtained greater importance in the research area of semantic web. The crux contribution of the paper is to present a framework known as DEKA for constructing and visualizing ontology for the web documents based on Computational Linguistics (rule-based model). The paper also addresses the working of some of the automatic ontology construction methodologies. The experimental data sets taken from the web corpus have been found to produce promising results for the performance metrics.