Title: Integration of Demographic, Behavioral, and Contextual Data in AutoGluon for Holistic Buyer Modeling
Authors: Mayowa Alonge
Volume: 9
Issue: 5
Pages: 26-34
Publication Date: 2025/05/28
Abstract:
With the era of data-driven decision-making, consumer behavior modeling requires the fusion of various data modalities for an exhaustive buyer profile. This research explores the incorporation of demographic, behavioral, and contextual data in AutoGluon, an open-source AutoML library, with the vision of developing an integrated buyer modeling framework. We present a pipeline that preprocesses heterogenous data sources, imputes missing values, and encodes complex feature interactions to train high-performance predictive models with minimal human tuning. With multi-modal ensemble learning and feature importance of AutoGluon, the framework facilitates robust buyer segmentation and propensity scoring. Empirical experiments demonstrate significant performance gains over unimodal models, particularly in dynamic e-commerce and personalized marketing settings. The findings demonstrate the value of contextual enrichment and behavior-demographic synergy in predictive buyer modeling, with specific emphasis on AutoGluon's effectiveness for large-scale deployment of AutoML for business intelligence in the real world.