The document describes a proposed association rule-based recommendation system using big data and the Hadoop framework. The system aims to recommend products to online shoppers based on the products currently in their cart. It extracts frequent itemsets and association rules from transaction data using Apriori algorithm implemented as MapReduce jobs. This allows the system to scale to large datasets. The results show it can provide accurate recommendations even for first-time users, solving the cold-start problem. However, it does not provide personalized recommendations compared to collaborative filtering systems.