Logistic Regression Validation Approach

Hi, I am using matlab to run the Logistic Regression. I need to validate the logistic regression model. My approach is I split my data into two groups. 85% of them is for training, and 15% of them is for testing. Now I have model from training data sets. I use the independent variables of test data sets as the inputs for the model, and I got the predicted outputs of the model. Then, I have to compare the predicted outputs and observed outputs. What kind of method here should I use to compare them? MSE? or what? Do you think my approach is right? I've read some paper about goodness of fit. I am not quite sure for this issue. Could you please give me some guide? Thank you.