Enabling Decision Support Through Ranking and Summarization of Association Rules for TOTAL Customers

Abstract

Our focus in this experimental analysis paper is to investigate existing measures that are available to rank association rules and understand how they can be augmented further to enable real-world decision support as well as providing customers with personalized recommendations. For example, by analyzing receipts of TOTAL customers, one can find that, customers who buy windshield wash, also buy engine oil and energy drinks or middle-aged customers from the South of France subscribe to a car wash program. Such actionable insights can immediately guide business decision making, e.g., for product promotion, product recommendation or targeted advertising. We present an analysis of 30 million unique sales receipts, spanning 35 million records, by almost 1 million customers, generated at 3,463 gas stations, over three years. Our finding is that the 35 commonly used measures to rank association rules, such as Confidence and Piatetsky-Shapiro, can be summarized into 5 synthesized clusters based on similarity in their rankings. We then use one representative measure in each cluster to run a user study with a data scientist and a product manager at TOTAL. Our analysis draws actionable insights to enable decision support for TOTAL decision makers: rules that favor Confidence are best to determine which products to recommend and rules that favor Recall are well-suited to find customer segments to target. Finally, we present how association rules using the representative measures can be used to provide customers with personalized product recommendations.

Geographic data-mining and extraction of association rules using the Apriori algorithm (Case study: Capital of Iranian provinces)

“…In addition to support and confidence metrics, other criteria have been proposed to measure the attractiveness of rules. Important metrics are as follows: 1) Lift [25] 2) Conviction [26] 3)Jaccard [27] 4) Leverage of Novelty [28] 5) Dependency Matrix [29].…”

Multi-Objective Recommendations and Promotions at TOTAL

“…We use a classical Item-based Collaborative Filtering (IBCF) approach [24] which calculates a similarity sim(i, j) between each pair of products i and j using cosine similarity. Our model accommodates other similarity functions and recommendation strategies such as association rules [3]. We choose IBCF for its better precision on our dataset [1].…”