An association graph of correlated values found with BigML's association rules (in collaboration with Geoff Webb) on the UCI mushroom dataset.
Move the slider to increase or decrease the number of rules used to build the graph. The highest quality rules (as defined by leverage) are included first. As the number of rules increase, the "min leverage" will drop indicating that weaker associations are being
| <!DOCTYPE html> | |
| <meta charset="utf-8"> | |
| <style> | |
| #rules { | |
| width: 960px; | |
| } | |
| #rule-table { |
A partial dependence plot of a BigML random decision forest built on the iris dataset.
A 2D partial dependency plot of a random decision forest trained to predict forest cover.
A 2D/3D heatmap of the forest cover dataset (100K sample).
Click and drag to zoom. Single click to zoom out.
A 2D/3D heatmap of the iris dataset.
Click and drag to zoom. Single click to zoom out.
A t-SNE inspired visualization of a BigML LDA model built from a subset of the UCI OpinRank dataset. Mouse over a topic to see the most influential terms.
| (define (merge-with* fn m1 m2) | |
| (reduce (lambda (result key) | |
| (let (v1 (get m1 key) | |
| v2 (get m2 key)) | |
| (assoc result key (if (and v1 v2) | |
| (fn v1 v2) | |
| (or v1 v2))))) | |
| {} | |
| (concat (keys m1) (keys m2)))) |
A t-SNE inspired visualization of a BigML LDA model built from a subset of the UCI OpinRank dataset. Mouse over a topic to see the most influential terms.
A visualization of a distribution encoded by a streaming approximate histogram.
t to toggle trimming some of the outliers from the distribution.r to toggle rounding populations for each bin.i to toggle the distribution interpolation mode.