CatorDogMachineLearningImageClassifierProject

Created and analyzed a machine learning project for determining if an image is a picture of a dog or of a cat. 1,170 images from the Kaggle data set were used for feature selection performed by a pre-trained Convolutional Neural Network to create a Cosine K-Nearest Neighbor classifier and a Cubic Support Vector Machine classifier. 25,000 images were then used for a second feature selection to create another Cubic Support Vector Machine since it was the better performing classifier. The entire machine learning workflow was performed including partitioning the data, feature selection, classifier selection, training, cross-validation and testing. The MATLAB Classification Learner App was used to help create the classifiers. This project was performed by two partners and myself for my Introduction to Artificial Intelligence elective course (CAP4630).