Experimental Report on Improved Generalized Mapping Algorithm in E-commerce Fraud Detection

Abstract

This experiment focuses on the field of e-commerce fraud detection, aiming to verify the performance advantages of the improved generalized mapping algorithm compared with the Markov chain. Simulated order data with 9 features were generated via Python, and both algorithms were used for processing. The experimental results, analyzed from aspects of time complexity and accuracy, show that the improved generalized mapping algorithm surpasses the Markov chain in detection accuracy, providing a more reliable solution for e-commerce risk control. In terms of processing efficiency, the Markov chain has a speed advantage for small-scale orders, while the improved generalized mapping algorithm is more efficient for large-scale orders.