Importance of error term in linear equation




Recently I have watched a you tube videos on linear regression and it is showing linear equation as y= a+bx+e (error term). Please help me to understand this error term, does this play a vital role in model development?

Thanks, Pravin



The aim of linear regression model is to find / predict the coefficients “a” and “b” to minimize the the presence of error term “e” in the predicted output. Here one has to keep a check that error term is only produced due to randomness of the process and there is absence of heteroscedasticity i.e. variances of the error term is equally distributed across zero. This will ensure that there is no pattern left in the error term which can be explained by the predictor variables and as a result you will get the best model output.