Generalized-Linear-Models.md

@@ -172,6 +172,8 @@ In summary, a generalized linear model can be expressed as:
 
 <img src="https://latex.codecogs.com/svg.latex?y_i&space;\sim&space;f(y|\mu_i,\phi)\\&space;\indent&space;g(\mu_i)=\eta_i\\&space;\indent&space;\eta_i=\sum_q{\beta_qx_{iq}}=\boldsymbol{\beta^{'}x}_i" title="y_i \sim f(y|\mu_i,\phi)\\ \indent g(\mu_i)=\eta_i\\ \indent \eta_i=\sum_q{\beta_qx_{iq}}=\boldsymbol{\beta^{'}x}_i" />
 
+If a GLM does not fit the data well, the considered distribution for the response variable may not be appropriate, the link function may not be appropriate, the linear predictor may not contain all relevant variables, or a combination of the mentioned problems. 
+