FP-CongQiaoben
Contents
Proposal
Group Members
Cong Qiaoben
Description
This project will explore the visualization of the adjacency matrix, in particular, the adjacency matrix of the match-up (win-rate) data. Instead of clustering the players directly using machine learning technique, this project will aim to improve the visualization by setting heuristics (evaluation) of the visual component. The final presentation will be an improved adjacency matrix with clear cluster as well as the algorithm that automatically generates it.
Project Progress Presentation
Progress Report
Literature Review
This project will primarily base on the following literature reviews
1. The history of the cluster heat map gives an extensive overview of the existing algorithms for clustering heat map, as well as common tools.
2. Seriation and matrix reordering methods: An historical overview shows the prior approaches for reordering matrices.
The problem of reordering matrix is in general NP-complete. Therefore, both reviews suggest that the ordering of the matrix bases a lot on the heuristics. However, the common heuristics (distance between rows and columns) might not work very well for the match-up data set.
Reference
1. M. Friendly. The history of the cluster heat map. The American Statistician, 2009.
2. I. Liiv. Seriation and matrix reordering methods: An historical overview. Statistical Analysis and Data Mining, 3:70–91, 2010
MileStone
1. Evaluation of existing algorithm. 5/25.
2. Try new heuristics & model. 5/27.
3. Try new visual encoding. 5/30.
4. Prepare final project poster presentation. 6/3
Project Plan
Final Deliverables
- Link to source code and executable
- Link to final paper in pdf form
- Link to final slides or poster
File:Cqiaoben poster.pdf File:Cqiaoben final.pdf File:Cqiaoben code.txt