Michael's Wiki

Guidelines

Define basic notation to explain the model. Establish a good staged plan.

Suggested length is 6 to 8 pages. Related work, details, but importantly: insight/analysis. Concisely show these results. Should be along the lines of a “mini” research paper.

If results didn't turn out well, or got negative results, that's ok. Important part is insight and the clarity of the writing.

Due on 8am Monday March 17th. Submit to EEE; no hardcopy required.

Software options:

• * General programming languages
• intended for final implementations
• * Scientific programming languages
• intended for prototyping and development
• * Data analysis packages
• intended for analysts with limited programming skills

Introduction to data mining

Exploratory Data Analysis *Visualization

Distance Measures Clustering

Unsupervised Learning *Dimensionality Reduction

The Project

Canceled for MLK Jr. Day

The Project Dimensionality Reduction

Predictive Modeling

Ranking Ranking Visualization Linear Regression

Recommender Systems

• * See Recommender Systems, Melville and Sindwhani or other recommended papers on the class syllabus

Recommender Systems Matrix Factorization Gradient Descent

Netflix Prize Text Analysis

Text Classification (http://www.ics.uci.edu/~smyth/courses/cs277/public_slides/text_classification.pdf)

• * Naive Bayes Classification
• * Multinomial Classification

Ranking Logistic Regression (see http://www.ics.uci.edu/~smyth/courses/cs277/papers/elkan_logisticregression_notes.pdf)

Cancelled for President's Day holiday

Topic Models

• * Background reading: David Blei's Topic Model webpage, the original LDA paper, the Steyvers LSA paper, Newman's distributed algorithms paper.

Review of Topic Models

Web Data PageRank

Web Data* Markov Chains

Computational Advertising

Multi-Armed Bandit Problem Logistic Regression Ranking Retrieved Documents for a Query * try (http://www.ics.uci.edu/~smyth/courses/cs277/public_slides/Web_advertising.pdf)

Class project demonstrations The Project