Comparison of Regression Techniques and Integration Models Using Machine Learning

Regression problems help extract the law from massive amounts of data to predict. Hence, regression algorithms have been a hot topic in machine learning research in recent years. In actual problems related to regression have complex internal and external factors, and different machine learning algorithms affect scalability and predictive performance. This research uses BP neural network, extreme learning machine, and support vector machine to examine the applicability of machine learning methods in regression problems. Comparing the single model and integrated model regression results of each machine learning algorithm shows its pros and cons. Finally, simulated experiments using four data sets show how well each machine learning method predicts regression. Research on machine learning algorithms and integration models appears to be viable and logical.