COMPLETE THE MATLAB CODE

The sigmoid function is used in logistic regression to calculate probabilities of belonging to one of two classes. The function, in its purest form 0(x) = -1 1 te- maps negative x to values less than 0.5 and positive x to values larger than 0.5. However, this form might not fit the data well. Suppose you wanted to predict if a student would pass an exam (1) or not (0) based on the number of hours spent studying x. In this case, the feature x cannot be negative, so everything would be predicted as 1 (pass). We can introduce trainable parameters a and b to give the sigmoid function more flexibility to fit the data: 0(x; a, b) = - 1+e-ax+b) The goal in logistic regression is to train the algorithm (with gradient descent) to find the a and b that best fit the data. The range of this function is still (0,1); the parameters only shift the function left/right and change its curvature. In this problem, you will try two combinations of a and b, and observe how they change the shape of the sigmoid function and the prediction it yields. Specifically, you will see how the prediction of test case x = 1.0 depends on a and b. You can submit your work as many times as you want. Try to get 100%. for ii = 1:2 %each of two cases if ii == 1 %first combination %you can change these parameters, but set them to their default values when submitting %defaults: a = 2.1; b = 0.5 a = 2.1; b = 0.5; else %second combination %defaults: a = 0.4; b = -2.6 a = 0.4; b =-2.6; end figure) %makes a new figure window x = linspace(-15, 15, 1000); %creates 1000 xvalues between -15 and 15 y = %YOUR CODE HERE. calculate the sigmoid of each x using a and b parameters x_test = 1.0; %we will assign this test data point to either class or class 1 y_test = %YOUR CODE HERE. calculate the sigmoid of the test case using parameters a and b. Similar to Line 13 threshold = %YOUR CODE HERE. What value of sigmoid function separates class from class 1? %---print some diagnostics, in this case, the class assignment fprintf(' When a = %.1f and b = %.1f,',a,b) if y_test