The feature-value paradox: Unsupervised discovery of strategic archetypes in the smartphone market using machine learning | ||
| Journal of Artificial Intelligence in Engineering Practice | ||
| Volume 2, Issue 2, November 2025, Pages 65-72 PDF (593.27 K) | ||
| Document Type: Original Article | ||
| DOI: 10.21608/jaiep.2025.420689.1024 | ||
| Authors | ||
| Chrisogonus K. Onyekwere1; Chinedu Kingsley Nwankwo2; Okechukwu J Obulezi* 1; Christiana I. Ezeilo1 | ||
| 1Faculty of Physical Sciences, Nnamdi Azikiwe University, Awka 5025, Nigeria | ||
| 2Faculty of Sciences, University of Abuja, 902101 Abuja, Nigeria | ||
| Abstract | ||
| This study employs unsupervised machine learning, a core branch of Artificial Intelligence (AI), and feature importance analysis to identify strategic archetypes in the smartphone market based solely on technical specifications. Moving be- yond traditional price prediction models, we analyze a comprehensive dataset to discover latent product strategies. Using K-Means clustering, we identify five distinct strategic archetypes, which we then analyze against price categories to re- veal both aligned and paradoxical positioning strategies. Our findings demonstrate that approximately 23% of devices exhibit a feature-value paradox, where premium specifications are not rewarded with premium pricing. Through permutation importance analysis, we quantify the feature importance driving each archetype. This research contributes to marketing science and engineering practice by offering a novel, AI-driven methodology for reverse-engineering product strategies, with direct implications for product portfolio optimization and competitive positioning in technology markets. | ||
| Keywords | ||
| Product positioning; Strategic archetypes; Artificial Intelligence; Machine learning; Smartphone market; Unsupervised learning; Feature-value paradox | ||
|
Statistics Article View: 15 PDF Download: 10 |
||
