Every other point determines which team

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: , line thickness, etc. be used to highlight pa\erns in the data? •  Data explora8on or debugging tool? (Iterate) •  Or final visualiza8on? •  Interac8on: e.g., selec8on or filtering •  •  •  •  •  h\p://eagereyes.org /techniques/parallel- coordinates 2 2/22/12 Radar Chart (a.k.a. web chart spider chart, star chart, star plot, cobweb chart, irregular polygon, polar chart, kiviat diagram) Today’s Class •  •  •  •  •  •  •  •  Readings for this Week Examples of High Dimensional Data Parallel Coordinates Data Clustering Principle Components Analysis (PCA) General Massive Data Visualiza8on Tips Next Week’s Readings Assignment 5 & Mid- Term Presenta8on From NASA: h\p://en.wikipedia.org/wiki/File:MER_Star_Plot.gif K- Means Clustering Clustering & Parallel Coordinates For a set of 2D/3D/nD points: 1.  Choose k, how many clusters you want (oracle) 2.  Select k points from your data at random as initial team representative 3.  Every other point determines which team representative it is closest to and joins that team 4.  The team averages the positions of all members, this is the team’s new representative 5.  Repeat 3-5 until change < threshold h\p://wanderinforma8ker.at/unipages/ParCoord/clustering_en.html How to do (K- means) Clustering Today’s Class •  Determine your distance func.on •  •  •  •  •  •  •  •  –  In spa8al datasets, o[en just be Euclidean distance •  Maybe also add in surface normal, etc. –  Rela8ve weigh8ng of different dimensions •  Especially tricky when units are unrelated convert to % of range •  Also problema8c when values are binary •  Finding nearest neighbors can be expensive –  Use a spa8al data structure Readings for this Week Examples of High Dimensional Data Parallel Coordinates Data Clustering Principle Components Analysis (PCA) General Massive Data Visualiza8on Tips Next Week’s...
View Full Document

This document was uploaded on 03/16/2014 for the course CSCI 4973 at Rensselaer Polytechnic Institute.

Ask a homework question - tutors are online