基于双塔模型的车货匹配方法

Vehicle-Cargo Matching Method Based on Two-Tower Model

车货匹配是网络货运平台的核心功能,车货匹配方法一直是人们关注的焦点。然而,研究中并没有重视大规模历史数据,且很少与深度学习的方法相结合。针对上述问题,本文利用双塔模型来解决车货匹配问题,通过对历史数据的处理与学习,预测司机点击货物的概率,通过某网络货运平台所提供的数据集进行实验。实验结果表明,双塔模型可以高效地处理数据,而且具有不错的准确性。

Vehicle and cargo matching is the core function of online freight platforms, and vehicle and cargo matching methods have been the focus of attention. However, the research has not paid attention to large-scale historical data and rarely combined with deep learning methods. To address the above problems, this paper uses the twin-tower model to solve the vehicle-goods matching problem by processing and learning historical data to predict the probability of drivers clicking on goods, and conducts experiments with the dataset provided by a web-based freight platform. The experimental results show that the twin-tower model can process the data efficiently and with good accuracy.