A Comparative Study of Linear and Nonlinear Machine Learning Algorithms in Quantitative Investment

In quantitative investing, multi-factor models are widely used, but the growing number of factors presents challenges for traditional models in factor selection and model structure. Machine learning algorithms, however, offer advantages in feature selection and return prediction, enabling multi-factor strategies to handle more data and achieve better performance. Linear and nonlinear machine learning algorithms have distinct feature selection mechanisms, each with its strengths and weaknesses in return prediction.