Resist the temptation to google for a solution. If you discover the pattern early, please resist the temptation to share the answer.

Project 6: What's in a Name? (Applied Theory & Practice)

Objectives

• Learn to undergo the data mining process
• Exercise simple feature creation
• Gain a deeper, hands-on understanding of some algorithms we've studied

Requirements

You might want to watch this (~20 min) first.

Given this list of names(csv, sqlite), can you discover what it is about them that results in each name being labeled as "winner" or "loser?" Using a data mining tool of your choice, can you build a model that can predict whether or not you are a "winner" or "loser?"

Winners in the dataset are labeled with a +. Losers, a -.

Write a brief paper that describes the process you undertook while trying to discover if patterns exist, and what those patterns are. Your paper should "tell a story": it must be clear to the reader exactly what actions you took and why. Include screenshots, summary stats and other measures of your results, where relevant.

Hints

There are multiple approaches to this problem, but I encourage you to:

• Explore the data. (How many are losers? Winners?)
• Use a data mining strategy via Weka or Knime (or an independent program) to see if a pattern exists.
• Implement a program that creates features based on the names. (eg, a boolean attribute representing whether or not the name rhymes with "snails.")
• Re run your data mining strategy upon this data.

Grading Criteria (1000 points)

You must have committed and pushed a copy of your paper, which effectively summarizes the data mining process you undertook.

Due Date

This assignment is due by midnight on Friday, October 7.

Resist the temptation to google for a solution. If you discover the pattern early, please resist the temptation to share answer.