CASH-GA-BO: A Hybrid Approach for Algorithm Selection and Hyperparameter Optimization using Genetic Algorithms and Bayesian Optimization

The use of automated machine learning [1]AutoML to solve problems in the real world is revolutionising. The selection of models and hyperparameter adjustment are the main topics of this paper’s overview of AutoML. The optimum machine learning model and parameter configuration are sought for through model selection and hyperparameter optimisation. The proposed AutoML system uses probabilistic reinforcement learning and prior knowledge to handle the problem of the expansion of the solution space. The paper illustrates the promise and difficulties of AutoML while also discussing existing approaches including [2]CASH, [13]Bayesian optimisation, and genetic algorithms. In general, AutoML simplifies the machine learning procedure, enabling academics and professionals to effectively address challenging issues.