Customizing an Inference System to Extract Knowledge from Data Repositories Using Concepts of Fuzzy

Algorithms are in abundance for extraction of information but some are not up to the expectation. It is due to the lack of effective mechanism in place to represent the required information. Fuzzy logic provides a sophisticated way of expressing the requirement with its membership functions, rule base and reasoning mechanisms. Membership function helps to flexibly specify the grade and extent of information needed and the rule base provides methods for combining them for effective representation of what is required. The reasoning mechanism in fuzzy logic helps to convert these representations into appropriate values so that the information can be classified according to our needs. In this research endeavour, a methodology has been delineated to describe the required information and a fuzzy model has been defined to extract the documents. It is flexible to change the required information and its extent using fuzzy logic. Using different types of rule sets how the extraction of information is being changed, has been discussed. Fuzzy inference system has been modelled to rank the documents in the order of the grade of information content. An algorithm has been provided to extract the information from rule sets. A generalized membership function has been delineated in this research work. By changing the parameter values in the membership function different membership functions can be initiated. By using two different sets of parameter values, two different membership functions have been modelled in this research work. These values combined with two different inference models namely, mamdani and sugeno models, totally four different fuzzy models have been implemented in this research work and the results are shown in the graphs at the end of this research work.