Save-Points Detection for Machine Unlearning in Multiple Linear Regression Models

Good machine learning models are evidently available in the literature. More research towards even improved versions of these machine learning models is still on going. However, the machine unlearning side of the coin has been ignored. The little that has been done focused on erasing all the data previously used by a machine learning model as a means of making the machine learning model forget the learnt trends. However, erasing all prior learning experiences in order to then re-learn from scratch is time consuming and expensive. In addition, there are high risk of datasets pollution in the process. This study investigated the detection of save points in multiple linear regression models. Results show successful detection of save points where major trend changes in the data are visually noted. This is a plausible model that will potentially improve the quality of machine learning, as well as its speed.