A Multilayer Perceptron based Neural Network that classifies an image between 10 distinct classes, this program uses a huge image dataset from CIFAR 10 and using 3 separate Neural Network models, classifies them into 10 distinct categories of objects. The difference betweeen the 3 models, are the various parameters that affect the classification like accuracy, losses, optimisation methods like Early Stopping and Dropouts.
A detailed Prediction Distribution table has been made for the 2nd Model and by minor tweaking of variables can be plotted for the other two models as well. An analysis of the F1 score and the recall and precision parameters have also been made at the end of the code. Images to be uploaded soon.