Recent Advances in E-Commerce System for Sale Prediction Using Triple Pseudo-Siamese Network with Giant Trevally Optimizer: A Systematic Review

Abstract

The rapid expansion of digital commerce platforms has generated vast amounts of transactional and behavioral data, creating significant opportunities for advanced analytics in retail decision-making. Sales prediction is a critical application that supports inventory optimization, demand forecasting, pricing strategies, and supply chain efficiency. Traditional forecasting methods such as regression models, ARIMA, and exponential smoothing have been widely used; however, they often fail to capture the nonlinear and dynamic patterns present in modern e-commerce data.Recent advancements in machine learning and deep learning have significantly improved forecasting accuracy. Techniques such as random forests, gradient boosting, convolutional neural networks (CNN), long short-term memory (LSTM), and transformer models effectively analyze complex temporal and behavioral patterns. Among emerging approaches, Siamese neural networks have gained attention for their ability to learn similarity relationships across heterogeneous data sources, enhancing predictive performance by capturing interactions between products, customers, and historical sales data. Furthermore, nature-inspired optimization techniques such as the Giant Trevally Optimizer (GTO) improve model convergence and parameter tuning. This review examines recent developments in e-commerce sales prediction, highlighting hybrid models integrating deep learning and optimization for scalable, accurate forecasting systems.