Project Proposal:Daniel and Sherry

Team Members

Daniel Mills

Shaghayegh (Sherry) Sahebi

Dataset

Russian movie social network data (Detailed information pending)

Project Ideas

We propose to explore the properties of a large social network tied with movie ratings and comments. We will investigate a variety of topics within this, using a mix of supervised and unsupervised machine learning methods.

Tasks

• Predict user ratings of movies based on their ratings of other movies, ratings made by there friends, and comments by them and friends.
• Detect hidden communities using both the existing social network as well as rating and comment information
• Predict evolution of the social network using interest similarity
• Predict popularity/revenue of movies based on all available information

Evaluation

Using cross-validation, we can compare predicted user ratings with actual ratings. Hidden communities are harder to evaluate, but can potentially be used as features in other tasks. Prediction of change in the network can be directly measured using recall and precision of new links predicted, and predicting popularity and revenue can be evaluated directly by comparing the predicted value to the actual.

Potential Methods

• Analyze data for correlation before doing anything else
• Graph clustering, such as spectral clustering
• Regression models
• Topic models, potentially hierarchical or with hidden variables