Abstract: The world is generating a massive amount of data. Therefore, there is a need for efficient methods to analyze and visualize enormous amount of valuable data being generated every day. Many methods are available for data mining, which extracts knowledge from data. But there is no method available which outperforms rest of them. Therefore, we have developed an algorithm based on a classic apriori algorithm and fp-growth algorithm to extract knowledge from data. We have used trie data structure to improve the performance by reducing the number of database scans. We have tested our algorithm on End of Day (EOD) data from November 2003 to August 2016 of Multi Commodity Exchange (MCX) of India. We found that our algorithm is faster than classic apriori algorithm.

Keywords: Improved Apriori Algorithm, Frequent Itemset Mining, Data Mining, Multi Commodity Exchange (MCX), Commodity Market