CHARM: Context-based Hierarchical Association Rule Mining for Analyzing Purchase Patterns

Title

CHARM: Context-based Hierarchical Association Rule Mining for Analyzing Purchase Patterns

Creator

Sahni, Pronnati; Ummesalma, M.; Parihar, Ruchi Singh

Description

Data mining is now an essential part of business intelligence, specially in the retail analytics, allowing companies to derive meaningful insights out of big volumes of transaction data. This paper uses Context-Based Hierarchical Association Rule Mining to study purchase behavior in Indian retail outlets through Apriori algorithm that helps to take effective decissions. The available literature primarily employs flat item association models and lacks contextual dimensions and profit-oriented outcomes of rules, which also creates an evident gap in the current research. The study combines various contextual aspects, including product category, sub-category, region, and state, to produce the multilevel association rules indicating the product relationship under different sales levels of the products following an hierarchy. The Lift and Conviction metrics are applied along with support and confidence to eliminate the coincidental patterns and make the rules in business reliable. Support-based filtering and a minimum threshold of confidence of 0.1 are used to determine separate patterns of co-purchase that are significant. In order to make business relevant, the level of profit is involved as a result which puts into emphasis rules which lead directly to financial performance. The findings show that context-enriched rules offer a better insight into customer buying behavior and retailers have the opportunity to identify profitable cross-selling opportunities that more traditional flat associated analysis might otherwise miss. The hierarchical structure allows improving interpretability through associating items with larger contextual properties, which will be useful in designing the promotion, product placement, and optimizing the regional strategy. Overall, this paper presents the combination of contextual and profit-driven parameters as a concept that can be used to provide a data-driven basis of strategic retail decision-making and sustainable competitive advantage. 2026 IEEE.

Source

2026 2nd International Conference on Advances in Intelligent Computing and Applications, AICAPS 2026;

Date

01-01-2026

Publisher

Institute of Electrical and Electronics Engineers Inc.

Subject

Apriori; Business Intelligence; Data Mining; Hierarchical Association Rule Mining; Market Basket Analysis; Purchase Patterns

Coverage

Sahni P., CHRIST(Deemed to be University), Department of Data Science and Statistics, Bengaluru, India; Ummesalma M., CHRIST(Deemed to be University), Department of Data Science and Statistics, Bengaluru, India; Parihar R.S., CHRIST(Deemed to be University), Department of Data Science and Statistics, Bengaluru, India

Rights

Restricted Access; Hardcopy may be available in the library

Relation

ISBN: 979-833159180-9;

Format

online

Language

English

Type

Conference paper

Identifier

https://doi.org/10.1109/AICAPS68631.2026.11453192

https://www.scopus.com/pages/publications/105036890783?origin=resultslist