Predictive modeling is an essential component within the health sector, as it serves to enhance the decision-making process and facilitate comprehension of the results for patients. Machine learning structures have been in existence for millennia, and it frequently appears that a significant number of algorithmic approaches lack accuracy and tend to provide erroneous predictions. In order to address this issue, our study proposes the implementation of optimization approaches to enhance the accuracy of algorithmic predictions. In this study, we are implementing Linear and Polynomial optimization approaches to Relational methods. We will afterwards evaluate the performance of these optimized Regression algorithms with their non-optimized counterparts. The evaluation of their efficacy will be based on the metric of Sum of Square Error (SSE). In the context of regression algorithms, it is generally preferred for the sum of squared errors (SSE) to be minimized, as this indicates a higher level of algorithm performance.