Optimization of Deep Learning Models for Dynamic Market Behavior Prediction

2025-11-24

AI Summary

This paper addresses the multi-horizon forecasting problem of SKU-level daily demand (or revenue) in e-commerce settings, supporting prediction horizons of 1, 7, and 14 days. We propose a hybrid deep learning architecture integrating multi-scale temporal convolution, gated recurrent units (GRUs), and time-aware self-attention to jointly capture local patterns, long-term dependencies, and temporal heterogeneity. The model explicitly incorporates dynamic market effects—such as holidays—to enhance robustness during peak periods. Training follows strict chronological splitting, with optimization via regression loss. Experiments on real-world e-commerce data demonstrate statistically significant improvements over ARIMA, Prophet, LSTM, and standard Transformer baselines across MAE, RMSE, and sMAPE metrics. Ablation studies and statistical significance tests confirm the individual efficacy and synergistic contribution of each component.

Technology Category

• Machine Learning: Hardware-aware ML
• Search and Optimization: Learning to Search
• Computer Vision: Multi-modal Vision

Application Category

• Economics, Online Markets and Human Computation: Sustainability of Web economics
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs

Abstract

The advent of financial technology has witnessed a surge in the utilization of deep learning models to anticipate consumer conduct, a trend that has demonstrated considerable potential in enhancing lending strategies and bolstering market efficiency. We study multi-horizon demand forecasting on e-commerce transactions using the UCI Online Retail II dataset. Unlike prior versions of this manuscript that mixed financial-loan narratives with retail data, we focus exclusively on retail market behavior and define a clear prediction target: per SKU daily demand (or revenue) for horizons H=1,7,14. We present a hybrid sequence model that combines multi-scale temporal convolutions, a gated recurrent module, and time-aware self-attention. The model is trained with standard regression losses and evaluated under MAE, RMSE, sMAPE, MASE, and Theil's U_2 with strict time-based splits to prevent leakage. We benchmark against ARIMA/Prophet, LSTM/GRU, LightGBM, and state-of-the-art Transformer forecasters (TFT, Informer, Autoformer, N-BEATS). Results show consistent accuracy gains and improved robustness on peak/holiday periods. We further provide ablations and statistical significance tests to ensure the reliability of improvements, and we release implementation details to facilitate reproducibility.

Problem

• Optimizes deep learning models for dynamic retail market demand forecasting
• Focuses on multi-horizon SKU-level daily demand prediction using retail data
• Addresses accuracy and robustness challenges in peak/holiday period forecasting

Innovation

• Hybrid sequence model with multi-scale convolutions
• Gated recurrent module for temporal dependencies
• Time-aware self-attention mechanism for forecasting