Final Project Websites

Recently, the course has become too big to handle individualized final projects. In past years, the final project has been an opportunity for students to explore machine learning in the context of a real-world data set. Projects could be whatever the students want, as long as it is focused on getting a machine learning system to learn. The links below connect to the final project websites submitted for Machine Learning EECS 349 over the years. There are a lot of interesting projects here. Have a look!

Fall 2013

Instruction-level Timing Error Prediction Model for 5-stage Pipelined ALPHA processor

Judgit - A Reddit popularity predictor

Pavlov's Arcade: Learning to play Breakout with Q-learning

Wiki Predict

MLB Hall of Fame Predictor

Diagnosis of Viral Infection using Gene Expression Profiles

Music Genre Classification

Texture Classification

Got the goods to go Greek?

Kinect-o-vision: Object recognition in 3D

Automatic Classification of Tunebot Sources

Fall 2012

Reinforcement Learning in Score Four

Content-Based SMS Spam Filtering

Stock Market Prediction based on Historical Data and Market Sentiment

Ranking search query results by sentiment

Ready Spam Detection in Twitter

Learning based Malicious Web Sites Detection using Suspicious URLs

coLabel

G. A. Walker - Robot Texas Ranger

Topic Analysis of Scientific Documents using Clustering

Classifying Success and Failure on Kickstarter

Political Framing

Twitter Retweet Prediction

The (mini) Book Genome Project

Fall 2009

Learning Subjective Image Concepts

A Na´ve Bayesian Approach to Evaluating Computer- Mediated Communication

CRUPS: Consumer Review Utility Prediction System

Automatic Wood Defect Classification

Stock Price Time Series Forecasting by Neural Networks

Trust Me, I Know You: Using Clustering to Predict Renter Film Selections in Netflix

Twitter-Amazon based book recommendation engine with weighted attributes

Data fusion for powered lower-limb prostheses

Optimal Parameters to Use for Classification of Paintings

Pandora RSS

Fall 2007

Genetic Algorithm based Network Intrusion Detection System

Comparing Approaches to Connect6: Temporal Difference vs. Minimax

Learning to Search a Melodic Metric Space

Score Following with HMMs and Polyphonic Audio Input

Dynamic Cache Resizing and Voltage Scaling using Machine Learning Algorithm

Using Neural Networks to Classify Paintings by Genre

Muscle Location Optimization Using a Genetic Algorithm 

Winter 2007

Learning to Detect Pitch Accents using a Naive Bayes Classifier

Face Recognition with Neural Networks

Sentiment Classification with SVMs

Rapid Facial Expression Classification Using Artificial Neural Networks

Image Spam Hunter

Using Neural Nets to Recognize Handwriting

The Effect of Region of Activity Measures on Automatic Audio-Visual Speech Recognition

Hummer Specific Learning to Enhance Query-by-Humming

Project GestureMap

Ergodic HMMs for Visual Speech Recognition