Current Projects (in Progress)
General Assembly capstone project, built recommendation engine based on user reviews of Anime from myanimelist.net using Keras/Tensorflow to predict user scores, TF-IDF Cosine pairwise distance, Latent Factors Cosine pairwise distance, and User Review cosine pairwise distance to generate possible shows to recommend.
Used the Riot Games API to scrape match data and player stats to create a model that would predict win/loss for a games based on characters chosen and individual player-skill with those characters. Compared XGBoost and Wide+Deep Neural net architecture for classifier accuracy.
Used Pipelines, several type of complex feature extraction to evaluate a number of different classifiers, including K-Nearest Neighbors, RandomForest, Log-Reg, SVM and a Neural Net classifier.