primaryobjects/unsupervised

Applying unsupervised learning using K-means clustering.

Unsupervised

Examples of applying unsupervised machine learning using K-means clustering.

Read the tutorial: Intelligent Topic Detection with Unsupervised Learning

Colors

Unsupervised learning is applied to a data set of randomly generated colors. The red, green, and blue values are used as features to categorize each color under a specific parent category.

For example, purple might be categories as Red or Blue. Likewise, Sky Blue would be categorized under Blue.

Cluster Categories

• Red
• Green
• Blue

Results

The following graphs show the results of clustering and categorizing colors by their red, green, and blue values.

1,000 Randomly Generated Colors

100 Randomly Generated Colors

3 Detected Clusters Within Colors

Assigning Colors to a Cluster

Viewing Colors Within Their Cluster

Predicting the Category for New Colors

The following three colors were used as new data for predicting the category for.

  red green blue     hex x        y group label
1 241    52   11 #F1340B 1 15807499     2   red
2  80   187  139 #50BB8B 2  5290891     3 green
3  34    15  194 #220FC2 3  2232258     1  blue

Exchange Traded Stock and Bond Funds (ETF)

Unsupervised learning is applied to a data set of exchange traded funds. The percentage values for "Year to Date", "1 Year", "5 Year", and "10 Year" returns are used as features to categorize each ETF under a specific parent category. Example code is provided in R and JavaScript.

Cluster Categories

• International
• StockBigGain
• Stock
• Bond
• SmallMidLargeCap

Results

The following output shows the results of clustering and categorizing ETF funds based on their percentage returns.

Training Set Category Results

Results

Test Set Category Results

Results

Results from JavaScript

Results

License

MIT

Author

Kory Becker

http://www.primaryobjects.com/kory-becker