It is a big task to provide the accuracy of discovered relevance features in text documents for describing user requirements. Classification of data is biggest issue in more text documents because they have large number of words and data patterns. Most existing popular methods are used by word-based approaches. Still, they have all suffered from the problems of relevance and uncertainty. Over the years, there has been pattern-based methods should perform better result than word-based methods in describing user requirements. But, how to effectively use large scale patterns remains a typical problem in text mining. To overcome this problem, Fuzzy Relevance Feature Discovery Algorithm (FRFDA), classification techniques have been developed for relevance feature discovery. It describes both higher level and low level features based on word patterns. It is also classifies words into categories and updates those word weights based on their relevance and dispensation in patterns. The experimentation result proves that, the proposed FRFDA is better than existing manual and automation methods. The data set Reuters-21578 shows that the proposed model significantly outperforms faster and obtains better extracted features than other methods.

This paper deals with adaptive rule based machine translation from English to Telugu. This approach is based on rule-based methodologies. If-then methods to select the best rules for target language in translation, Probability based appropriate word selection for a given sentence and rough sets to classify a given sentence are the approaches used in this technique. Set of production rules of English and Telugu, Training set and Dictionary for both the languages are developed for this purpose. User gives and input, which is an English sentence. The given input sentence is then tokenized into individual words. These words are tagged with their respective parts of speech. All other words that are not found in the pre-defined database are tagged using grammatical rules that are formulated. Using these POS tags, the respective word translations are retrieved from the database. These individual words are then concatenated to form a sentence that is the result of user's input.