We used the admission data and tried to use the logistic regression with some of the features in order to predict whether a student with a given feature set (vector) will be admitted or not (the label). We fitted the regression, set SGD parameters (you should experiment), and ran the API. We then output intercept and model weights for the regression coefficients. Using the model, we predicted and output some predicted values for visual inspection. The last step was to output MSE and RMSE values for the model. Please note that this was for demonstration purposes only and you should use evaluation metrics demonstrated in the previous chapter for model evaluation and the final selection process. Looking at SME and RMSE, we probably need a different model, parameter settings, parameters, or more data points to do a better job.

The Signatures for this method constructor are as follows:

newLogisticRegressionWithSGD()

Defaults for Parameters:

• stepSize= 1.0
• numIterations= 100
• regParm= 0.01
• miniBatchFraction= 1.0