Machine Learning

Machine Learning

CodyM_Assignment1.pdf

Prediction

Code

I applied 3 regression models (OLS, Ridge, and Lasso) to marketing data to determine the most efficient ad spend to boost revenue.

CodyM_Assignment2.pdf

Classification

Code

I applied 4 classification algorithms to breast cancer tumor data to determine the best classifier of malignancy: Logistic Regression, Decision Tree, Naive Bayes, and Support Vector Machine (SVM).

CodyM_Assignment3.pdf

Cluster Analysis

Code

I applied the K-means algorithm to credit card spend data to find 4 distinct groups of customers that differ empirically and meaningfully from each other.

Deep Learning

CodyM_Assignment4.pdf

Classification

Code

Returning to the breast cancer dataset, I applied a Multilayer Perceptron (MLP) and sequential Deep Neural Network (DNN) to compare performance against metrics and thresholds important to medicine.

CodyM_FinalProject.pdf

Sentiment Analysis

Code

I compared the performance of Naive Bayes (NB) and a Recurrent Neural Network (RNN) on a consumer review data set from Yelp to determine sentiment.

automated-image-captioning-final.pdf

Image Captioning

Code

I applied pretrained Deep Learning models to generate captions for previously unseen photographs. Results were far below human capability; however, they were similar to state-of-the-art captions from the timeframe of the models publication in 2015.