Title: Refining mobile price predictions: leveraging regulated classification and quadratic discriminant analysis for enhanced accuracy

Authors: Chao Sheng Han

Addresses: Zhejiang Yuexiu University of Foreign Languages, Shaoxing, China

Abstract: Future mobile phone prices are forecasted using historical data and selected relevant features. Trends are analysed by brand and statistical parameters through machine learning methods. This supports consumers in making informed purchasing decisions and enables businesses to develop more effective pricing strategies and manage inventories efficiently. Accurate forecasts enhance market responsiveness, drive competition and streamline operations. In this study, Quadratic Discriminant Analysis (QDA) and Naive Bayes Classifier (NBC) were employed for price estimation. To optimise these models, two metaheuristic techniques - Manta Ray Foraging Optimisation (MRFO) and Escape Bird Swarm Optimisation (EBSO) - were applied. MRFO simulates manta ray foraging behaviour for parameter selection, while EBSO mimics bird escape behaviour for fine-tuning. These advanced approaches significantly improved forecasting accuracy and robustness. The QDA model optimised via EBSO, referred to as the QDEB model, achieved superior performance with an accuracy, precision and recall of 0.971, and an F1-score of 0.970.

Keywords: mobile price; classification; Naive Bayes classification; quadratic discriminant analysis; metaheuristic algorithm.

DOI: 10.1504/IJIPT.2026.156310

International Journal of Internet Protocol Technology, 2026 Vol.19 No.3, pp.156 - 170

Received: 26 Apr 2025
Accepted: 11 Jun 2025
Published online: 10 Sep 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article