The content introduces XRMDN, an Extended Recurrent Mixture Density Network, designed to forecast probabilistic demand in high-volatility Mobility-on-Demand systems. Traditional forecasting methods often overlook uncertainty in demand projections, especially in scenarios with high and dynamic volatility. XRMDN addresses these challenges by leveraging a sophisticated architecture that incorporates endogenous and exogenous data to enhance forecasting precision. The model outperforms existing benchmark models in various metrics, particularly excelling in high-demand volatility contexts. The paper includes a comprehensive experimental analysis using real-world MoD datasets, showcasing the effectiveness of XRMDN in enhancing operational efficiency and customer satisfaction.
Ke Bahasa Lain
dari konten sumber
arxiv.org
Wawasan Utama Disaring Dari
by Xiaoming Li,... pada arxiv.org 03-06-2024
https://arxiv.org/pdf/2310.09847.pdfPertanyaan yang Lebih Dalam