Hybrid-AutoML System Development

Abstract

This chapter presents the Hybrid-AutoML system requirements, design materials, model algorithms, and model design, which encompasses the design goals, architecture (a three-layered architecture), components, and characteristics of the Hybrid-AutoML toolkit developed in this research for automatic mode and model selection on single or multi-varying datasets. The mode components, decision learning and AutoProbClass unsupervised algorithms, and application API are described. The testing and evaluation of the model is conducted by two case studies.