Research Domain Selection plays an important role for researchers to identify a particular document based on their discipline or research areas. This paper presents a framework which consists of two phases. In the first phase, a word list is constructed for each area of the research paper. In the second phase, the word list is continuously updated based on the new domains of research documents. Primary area and Sub area of the documents are identified by applying pre-processing and text classification techniques. Naive Bayes classifier is used to find the probability of various areas. An area having the highest probability is considered as primary area of the document. In this paper text classification procedures is condensed as that are utilized to arrange the content archives into predefined classes. Based on the performance analysis, it has been observed that the obtained results are efficient when compared to manual judgement.

Text Summarization is a process that converts the original text into summarized form without changing the meaning of its contents. It finds its usefulness in many areas when the time to go through a large content is limited. This paper presents a comparative evaluation of statistical methods in extractive text summarization. Top score method is taken to be the bench mark for evaluation. Modified weighing method and modified sentence symmetric feature method are implemented with additional characteristic features to achieve a better performance than the benchmark method. Thematic weight and emphasize weights are added to conventional weighing method and the process of weight updation in sentence symmetric method is also modified in this paper. After evaluating these three methods using the standard measures, modified weighing method is identified as the best method with 80% efficiency.