Comparative Analysis of Quantitative Predictive Models Utilizing Machine Learning for Business Decision Making

Abstract

Multivariate analysis is a critical decision-making tool as it allows business managers to understand and predict complex behavior. This is an outline of how various quantitative approaches under the Machine Learning paradigm contribute towards optimizing business decisions. The research sought to evaluate, with measurement of accuracy, the results of the predictive model developed by eight quantitative techniques applied with Machine Learning algorithms executed in Python with the aim of finding the best model to adequately estimate the output variable: price. In data processing, they were loaded, cleaned and then divided into two categories: 70% for training and 30% for testing. The models used in the comparative study were: multiple regression, Ridge regression, Lasso regression, Decision Tree, Gradient Boosting, Random Forest, Support Vector Regression and Neural Networks. Benchmark performance, according to measures of accuracy such as mean square error, mean absolute error, and R-squared, showed that the best fits were for the model developed using Gradient Boosting with an R² of 0.8175, followed by the Random Forest model with an R² of 0.7889. The two models mentioned above were the most stunning, as they provided the most ideal key indicators for business decision making.