Abstract
Abstract : The emerging field of data science is dedicated to the analysis of vast datasets through diverse methodologies, aiming to enhance data comprehensibility. A solid grasp of statistics, linear algebra, and optimization is essential for a profound understanding of data science. The pursuit of model performance improvement involves the optimization of parameters and hyperparameters, seeking the most effective values for achieving desired outcomes. This study endeavors to introduce accessible algorithms, both new and established, emphasizing ease of implementation and comprehension. The focus extends to appreciating theoretical analyses and discerning which algorithm suits specific data science challenges. The objectives encompass the identification of various optimization problems, linking them with applicable solution methods, and the manual application of optimization techniques to solve modest problems. Additionally, the study delves into discussions on interpreting the sensitivity of optimization solutions to changes in parameter values.