Parameter Tuning Method Analytics (PaTM) for Different Datasets Using Classification Models

Abstract

In machine learning many classification models, have a range of parameters that may strongly affect the predictive performance of the model induced by them. Hence, it is recommended to define the values of these parameters using the optimization techniques. While these techniques usually converge to a good set of values, they typically have a high computational cost, because many candidate sets of values are evaluated during the optimization process. It is often not clear whether this will be a given results in parameter settings that are significantly better than the default settings. When training time is limited it may help to know when this parameter should definitely be tuned. Hence, in this thesis learning method has been used to predict when optimization techniques are expected to lead models whose predictive performance is better than those obtained by using default parameter settings. The parameter tuning method has also been utilized to improve the performance of classification models and reducing the error rate along with overall computational costs. We evaluate the proposed method on different datasets by selecting six datasets. The performance of parameter tuning framework is being evaluated and results show that the parameter tuning framework outperforms than other existing techniques.