Development of Innovative Machine Learning in Text Document Relevance Ranking

Abstract

With the flooding of Publications in the present era of Information Revolution, sophisticated add-on tools can be seen being rendered in order to evolve the Scholarly Communications towards the Digital Collections. The construction of Automated Digital Libraries has grabbed widespread popularity over recent years, rendering the value added services that suit to the tailored user requirements organizing text material in a finite sequence of topics and sub-topics. This can facilitate an end-user, fast and easy access to the course content of one s own interest. The context oriented information retrieval has always been based on some or the other explicit ontologies viz. Hierarchical thesaurus or subject oriented control dictionaries or vocabularies. The emphasis is on the Ontologies in Implicit form suitably extracted from the given text documents, upon which the organizing and retrieval tasks are to be carried through. newlineIn the pursed research, machine learning is incorporated to analyze the soft copies of the syllabi that can be assumed to be a part of Digital Libraries on the relevant subject(s), through congenial text mining techniques. The research focuses upon design of a system (tool) to rank text documents available in machine-readable format by analyzing them upon softcopies of the syllabus content, through congenial content filtering techniques. Document representation and modeling the text input too plays a significant role in accessing the filtered fragments yielding high performance accuracies. Also, there should be rational logic steered to direct the term-searches in shortest possible execution times from voluminous text corpora put to mining experiments. For this, the tool has been implemented into two phases: The first was the paragraph identification process and deciding page-vicinities of relevant text. For this, the paragraph-tagging process justifies itself to be a prelude step for identification of target and their relevant pages. In the subsequent phase, the semantic content filtering needed co