Genetic based method for Mining Association Rules from Text

Abstract

In general, text mining is the process of discovering very important and unknown information from a semi-structured or unstructured data set, while data mining deals with structured data. Where natural language processing technology was used: coding, stop word deletion, derivation and indexing of the document, convert text into structured form, and in order to extract correlation rules, we used the genetic algorithm (GA) to solve the correlation rule problem by overcoming the limitations of using the priori algorithm (a coding system, storage capacity). Were the estimates are based or based on a recommendation to use the individual variable length within the population. The rules for extracted relationships contain the essential features included in a document compilation. The proposed system aims to solve the problems facing traditional techniques as it works to address two main issues: Computational complexity, where the number of rules increases greatly with the number of elements in the database. Second, the rules related to interests must be collected within the place of creation of the rules that usually suggest a higher algorithm. The use of a genetic algorithm to find a relationship in the unorganized document between keywords is powerful and very effective in solving this problem because it effectively cuts the search space for threshold based consistency measurements.