Verification of a New Machine Learning SiNG in Classification Problems by a Large-scale Benchmark Problems
Mina Arakaki, Chikako Dozono, Hana Hebishima, Shin‐ichi Inage · 2023
This paper aims to propose and validate a new machine learning for classification problems. Classification problems are important in engineering applications of machine learning, including pattern recognition. Neural networks and other classification problem evaluation methods already exist and have been well validated. We propose a new classification method: SiNG, which has a simple configuration and a much lower computational load than neural networks. This paper applies the proposed SiNG to benchmark problems that consists of large input parameters and is of practical importance, and performs further verification of the SiNG. As a result, we confirmed the effectiveness of the SiNG even in large-scale benchmark problems. Keywords: Optimization algorithms, Machine learning, Classification problems, Monte Carlo method