Optimized Apriori Algorithm Using Matrix Data Structure

Association rule learning in data mining is a favored and well researched method for interesting relations between variables in large databases. Association rule of data mining involves picking out the unknown interdependence of the data and discovering the rules between those items. There are several algorithms for finding frequent item sets and mining comprehensive association rules from databases. In classical Apriori algorithm, algorithm may generate abundant number of candidate generations but traditional algorithms are not efficient. Every time Algorithm scans and judge about the candidate generation and frequency of item sets. This results into high frequency in querying, so amount of resources spent in terms of time or space is huge and decreases the efficiency of Classical Apriori Algorithm. To outcast these drawbacks we have proposed an optimized algorithm in this paper with an aim of minimizing the mundane and space complexities by reducing the database scans to one by generating impressed data structure bit matrix(bit_matrix) and by reducing superfluous computations for extracting regular item sets using top-down approach.