Optimizing Field-Aware Factorization Machine with Particle Swarm Optimization on Online Ads Click-through Rate Prediction

摘要

Online advertising industry is grow larger along with the increasing numbers of internet users. To make ads industry to be more efficient, prediction model for ads' click-through rate is needed. In this research, Field-aware Factorization Machine (FFM) is going to be optimized using Particle Swarm Optimization (PSO) on FFM parameters to increase the accuracy of the FFM. In this research, FFM and PSO-FFM is compared with accuracy and execution time. Our experimental results show PSO can increase FFM performance.