A linear regression model assumes that the underlying true relationship between the response variable and each predictor is [ Select ] . Variable selection techniques such as [ Select ] , regularization approaches such as [ Select ] , and dimension reduction techniques such as [ Select ] can improve the model performance when the assumption is met. Simple extensions of linear models like polynomial regression and step function as well as splines, local regression, and generalized additive models can maintain some level of [ Select ] while relaxing the assumption