The single biggest advantage of a linear model is that is simple. From there, there are mainly disadvantages and limitations.
- Linear model implies linear relationships.
A linear model assumes that the independent variables explain the dependent one in linear way. Using linear model, would either disregard some patterns or force us to execute complicated transformations to reach a linear representation.
2. Data must be independent
In 95 % of the linear models conducted in practice. Most linear models assume that the variables in the models are not collinear. We observe multicollinearity or the math behind the model estimation brakes.
3. Outliers
Since linear models assume linearity, having values that are too big or too small regarding any feature may be devastating for the model. All point are expected to be close to some line, which as you can imagine unrealistic. To deal with that we often complicated the linear model in ways that practically make it behave like a non linear one.
