Apriori based market basket analysis on hadoop

Author: 
Yugandhara Bhosale, Siddhesh Chendvankar, Anmol Wadekar and Nirav Dedhia

Data mining is one of the most important aspect in terms of business growth. Results generated from data mining is valuable which helps business grow. Finding frequent itemset is the key feature in data mining which has helped many business to grow immensely. In this paper based on Hadoop MapReduce Model we have implemented Apriori algorithm for finding the most frequent itemset and also for the recommendation purpose. Apriori algorithm select the pattern that occurred most and recommends next item accordingly which is now a days used in many e-commerce website to increase sales. MapReduce is a parallel data processing model. This model helps to express many computational problems. MapReduce model consists two functions Map function and Reduce function. In Map function each word is considered as separate string and is tokenized(value 1 is assigned).Reducer function will check the number of times that word appeared and will produce an output consisting that word and number of times it appeared in total.

Download PDF: