Research on the Construction of E-Commerce Market Trend Prediction Model Based on Big Data Analysis

In this study, a model of e-commerce market trend prediction based on big data analytics, Gated recurrent Unit (GRU) and Particle Swarm Optimization (PSO) is proposed. Firstly, the big data processing framework Spark was used to collect multi-source data from the e-commerce platform, including user behavior and sales records, and Spark SQL was used for data cleaning and preprocessing. Next, the GRU model is used to process time series data to capture the time dependence of the e-commerce market for trend prediction. GRU effectively controls the forgetting and memory of historical information through update gate and reset gate, so as to improve the prediction accuracy. To further enhance model performance, this study incorporates the PSO algorithm to fine-tune the hyperparameters of the GRU, such as the learning rate and the number of hidden layer nodes. By iteratively optimizing the particle swarm, the model can efficiently search for the best parameter combination, thereby improving its predictive accuracy. Ultimately, the effectiveness and precision of the proposed method in forecasting e-commerce market trends are validated through model evaluation. The model combines the big data processing ability of Spark, the time series modeling advantages of GRU and the global optimization ability of PSO, which is suitable for market trend prediction in complex e-commerce environment.