Rutgers University Python Cluster Training Dataset Project Use Python as programming language. You will be provided a training dataset (60k) and a test dat

Rutgers University Python Cluster Training Dataset Project Use Python as programming language. You will be provided a training dataset (60k) and a test dataset (10k) of handwritten digits. (http://yann.lecun.com/exdb/mnist/)

Question #1:

-Cluster training dataset with K=10 and assign a class to each cluster based on majority vote

-Compute the training accuracy. Use each cluster’s centroid as the anchor point and obtain the centroids

-Assign image to nearest centroid and test accuracy

Question #2:

-Build 2 non DNN based classifiers using all pixels as features for handwriting recognition. Use 2 of these techniques: Logistic regression, SVM, Decision tree, Random forest

-For each technique: describe technique, pros/cons, results

Question #3:

-You will be given a dataset of neural network and a python file for partial code

-Implement CNN (Convolutional Neural Network). Finish the Python code and report accuracy

Question #4:

-Hand write 5 styles of digits 0 to 9 on paper. Take a picture of each (50 total).

-Convert images to MNIST format

-Run ML models from question 1-3 and report accuracy

-Accuracy results of the 50 images should be shown in a table for each of the models