Thank you for coming here Purpose of Experiment
- Slides: 39
Thank you for coming here!
Purpose of Experiment n n Compare two visualization systems. You will play with one of them.
What will you do? n n n Learn a multidimensional visualization system; Use it to find features of a data set and record your result; A quick after-experiment feedback.
Schedule First, I will present. . . Multidimensional data Hierarchical Parallel Coordinates Brushing Feature finding Introduce the visualization system n
Schedule Then, You will do. . . Experiment: -Find features of a given data set using the visualization system -Record the features you find n Fill feedback form.
Outline n n Multidimensional Data How to represent multidimensional data n n n Parallel Coordinates Hierarchical Clustering Hierarchical Parallel Coordinates Brushing Operation Feature Finding
Multidimensional Data Example: Iris Data Scientists measured the sepal length, sepal width, petal length, petal width of many kinds of iris. . .
Multidimensional Data Example: Iris Data
Outline n n Multidimensional Data How to represent multidimensional data n n n Parallel Coordinates Hierarchical Clustering Hierarchical Parallel Coordinates Brushing Operation Feature Finding
Parallel Coordinates One-Dimensional Data (1. 6) 1 2
Parallel Coordinates 4 -Dimensional Iris Data Set
3. 5 5. 1 1. 4 0. 2
Outline n n Multidimensional Data How to represent multidimensional data n n n Parallel Coordinates Hierarchical Clustering Hierarchical Parallel Coordinates Brushing Operation Feature Finding
Hierarchical Clustering Hierarchical Cluster Tree A cluster tree
Hierarchical Clustering Mean, Extent y P 1( 3, 6) p 2( 5, 5) Mean Point of C 1 = P 1 (P 1+P 2)/2 = (4, 5. 5) P 2 C 1 x Extent of C 1: x: [3, 5] y: [5, 6]
Outline n n Multidimensional Data How to represent multidimensional data n n n Parallel Coordinates Hierarchical Clustering Hierarchical Parallel Coordinates Brushing Operation Feature Finding
Outline n n Multidimensional Data How to represent multidimensional data n n n Parallel Coordinates Hierarchical Clustering Hierarchical Parallel Coordinates Brushing Operation Feature Finding
Brushing - Highlighting part of the clusters to distinguish them from the other clusters.
Outline n n Multidimensional Data How to represent multidimensional data n n n Parallel Coordinates Hierarchical Clustering Hierarchical Parallel Coordinates Brushing Operation Feature Finding
Feature Finding Feature - Anything you find from the data set. Cluster - A group of data items that are similar in all dimensions. Outlier - A data item that is similar to FEW or No other data items.
Other features You can record anything else you find with the help of the visualization system.
