An ID3 GA Bass Collaborative Framework for Cross Border Market Diffusion Prediction

Abstract

Accurate forecasting of international market diffusion is essential for strategic decision-making in the New Energy Vehicle (NEV) sector. Traditional ID3 and Bass models face limitations in handling heterogeneous consumer behavior and dynamic policy shocks. This paper proposes a collaborative model that integrates an improved ID3 decision tree for consumer segmentation with a genetic algorithm-optimized Bass diffusion model. The ID3 component refines classification accuracy by mitigating multi-value bias, while the GA dynamically calibrates Bass parameters under policy interventions. Experiments using cross-country NEV data demonstrate that the proposed framework achieves a stable G-mean of 0.89 ± 0.03, with MAE and RMSE reductions exceeding 30% compared to benchmark models including deep reinforcement learning and optimized random forests. The model effectively quantifies the impact of consumer heterogeneity and subsidies on market diffusion, offering a decision-support tool for firms to design adaptive marketing and policy strategies. Beyond the NEV context, the framework can be applied to other domains requiring predictive control of diffusion processes under external shocks.