Do part C only

Data used in Problems 1 and 2 X1train and X2train are training data points in Class 1 and 2, respectively. Xtest are testing data points to be classified into Class 1 and 2 . In the following expression, two-dimensional data points are represented in columns: X1train=[23324432545328460065]X2train=[716915101311119171313914111511141012]Xtest=[40102.51012.5] (a) (20 marks) Carry out LDA to classify the data points provided above to Class 1 and 2 . Note that all calculations in Part (a) should be done by hand. The three steps involved are reproduced below for easy reference: Step 1. Apply LDA to project high-dimensional input vectors into one dimension. Step 2. Estimate , on the y-space for each of the two classes. Step 3. Place threshold on y for classification based on decision theory. Assume p1=0.5. (Hint: Make use of the results obtained in the tutorial). n particular, show the steps in your computation of the following items: i. S1,S2,Sw,SB,w,Y1,Y2 in Step 1 . ii. 1,1,2,2 in Step 2 . iii. The classification threshold obtained in Step 3. (Hint: Make use of the results obtained in the tutorial). b) (20 marks) Write a Python function called LDA to verify your calculation in Part (a). The function should have a header: def IDA(X1, X2) where X1 and X2 are arranged in columns. In the function, codes should be written to (1) compute Item i, ii and iii in Part (a), (2) plot of all the data points with the projection vector w superimposed and (3) f(yCk) for k=1,2 and (4)f(Cky) for k=1,2. The script should be submitted as an individual file named IDA.py and the codes written in the script should be copied and pasted to your report. (c) (10 marks) Write scripts in a cell of the Jupyter Notebook file named Problem2 . ipynb. In this script, you should: i. Train the model using the LDA function written in Part (b) using the training data. The value of each of Items i, ii and iii listed in Part (a) should be made to display on the command window. These values should be copied and pasted to the report. All the plots generated in Python should be included in the report. ii. Classify the test data points into Class 1 and 2