Solving Regression Problems using Haar Wavelets and Application in Prediction

The main purpose of this paper is to generalize our previous research papers on data science and machine learning models and then show that Haar wavelet basis is a powerful tool for solving regression problems in data science. In order to show this, we train two regression algorithms based on polynomial basis and Haar wavelet basis to predict a continuous target (car price). Among the wavelet families like Spline and B- Spline wavelets, the Haar wavelet, which is the simplest and easy to implement. The results of the comparison of the models show that the wavelet basis is more efficient and can make a much better prediction.

Keywords:

Machine learning , Data science , Haar wavelet , Regression algorithms