This document discusses estimating parameters for exponential distribution models using least-squares methods and optimization techniques, specifically focusing on the application to medical data from leukemia patients. It compares various optimization methods, including Nelder-Mead, Hooke-Jeeves, and quasi-Newton methods, assessing their effectiveness in achieving parameter estimates from Kaplan-Meier survival estimates. The findings indicate that both sets of optimization methods yield similar parameter estimates, emphasizing the robustness of the least-squares approach in statistical modeling.