An Adam-adjusting-antennae BAS Algorithm for Refining Latent Factors

TL;DR

This paper introduces the A2BAS algorithm, combining Adam optimization with BAS to refine latent factors in high-dimensional matrices, effectively addressing premature convergence in PSO-based latent factor analysis.

Contribution

The paper proposes a novel sequential A2BAS algorithm that refines latent factors in high-dimensional matrix analysis, improving upon PSO-incorporated LFA models.

Findings

Effectively solves premature convergence in latent factor extraction.

Demonstrates improved accuracy on real high-dimensional matrices.

Outperforms existing PSO-based methods in experiments.

Abstract

Extracting the latent information in high-dimensional and incomplete matrices is an important and challenging issue. The Latent Factor Analysis (LFA) model can well handle the high-dimensional matrices analysis. Recently, Particle Swarm Optimization (PSO)-incorporated LFA models have been proposed to tune the hyper-parameters adaptively with high efficiency. However, the incorporation of PSO causes the premature problem. To address this issue, we propose a sequential Adam-adjusting-antennae BAS (A2BAS) optimization algorithm, which refines the latent factors obtained by the PSO-incorporated LFA model. The A2BAS algorithm consists of two sub-algorithms. First, we design an improved BAS algorithm which adjusts beetles' antennae and step-size with Adam; Second, we implement the improved BAS algorithm to optimize all the row and column latent factors sequentially. With experimental results on two real high-dimensional matrices, we demonstrate that our algorithm can effectively solve the premature convergence issue.