MACHINE LEARNING

a. A machine learning expert has developed a classifier for a telecom company for detecting whether a cell-phone network base station is faulty or not based on training data provided by the company. He evaluated the classifier on a test set. Here is the confusion matrix: Predicted Not Faulty 5 965 Truth Faulty 10 20 Faulty Not Faulty 1) How accurate his classifier performs? 11) Which class is in majority and how well this classifier performs for the majority class? b. Figure. 1 is depicting training set comprising of two categories Red and Blue and two features Xi and X2 Figure 1 i) If you are asked to train a classifier using logistic regression and training set, will you use feature scaling before modelling? 11) Which degree of polynomial terms and parameters (B/O) would you prefer in the above case