ONLY MATLAB ONLY ONLY MATLAB #MATLAB

Write a function that will predict the unknown class label using the k-NN classifier algorithm. You can assume there will be two classes. Use the Euclidean distance measure. The function should accept as input: . the training set (the set including those instances with known labels, e.g., a 30x2 matrix for 15 sardines and 15 tuna fish, d=2), the label vector (the class labels of the training instances, e.g., a 30x1 vector), the instance with the unknown label (e.g., a 1x2 vector), the value of k, . . and give as the output: . the predicted class label of the instance. Write a script that will call this function: . . to predict the class labels of all the instances given in the validation set, to calculate TCC per given k value, to plot TCC vsk, to print (on the command window): the k value with the maximum TCC, and the corresponding the confusion matrix. . O Sweep over the values of k from 1 to 35 (Hint: k should be an odd value when you have two classes). Notes: You may decide your own way of writing your script/function. Correct outcome and efficiency should be your target. The training and validation sets for this project are given in the mat-file named DATA.mat. The third columns have the class labels. The data is arbitrary, i.e., the two dimensions are not the length and the width of sardines and tuna fish :) We can talk about both the training error and the validation error in classification. Accordingly, one can build two confusion matrices: one for the training set and the other for the validation set. Here, you are asked to build the latter: the one for the validation set only. We do not consider the instances in the validation set "neighbors to each-other as implied by the name of the algorithm. The training instances are the neighbors (of the validation instance) to look at, so to speak. . . Write a function that will predict the unknown class label using the k-NN classifier algorithm. You can assume there will be two classes. Use the Euclidean distance measure. The function should accept as input: . the training set (the set including those instances with known labels, e.g., a 30x2 matrix for 15 sardines and 15 tuna fish, d=2), the label vector (the class labels of the training instances, e.g., a 30x1 vector), the instance with the unknown label (e.g., a 1x2 vector), the value of k, . . and give as the output: . the predicted class label of the instance. Write a script that will call this function: . . to predict the class labels of all the instances given in the validation set, to calculate TCC per given k value, to plot TCC vsk, to print (on the command window): the k value with the maximum TCC, and the corresponding the confusion matrix. . O Sweep over the values of k from 1 to 35 (Hint: k should be an odd value when you have two classes). Notes: You may decide your own way of writing your script/function. Correct outcome and efficiency should be your target. The training and validation sets for this project are given in the mat-file named DATA.mat. The third columns have the class labels. The data is arbitrary, i.e., the two dimensions are not the length and the width of sardines and tuna fish :) We can talk about both the training error and the validation error in classification. Accordingly, one can build two confusion matrices: one for the training set and the other for the validation set. Here, you are asked to build the latter: the one for the validation set only. We do not consider the instances in the validation set "neighbors to each-other as implied by the name of the algorithm. The training instances are the neighbors (of the validation instance) to look at, so to speak