Skip to content

Baylex/Cryptocurrencies

Repository files navigation

Cryptocurrencies

Unsupervised Machine Learning and Cryptocurrencies

Purpose:

A client asked for a list of tradable cryptocurrencies and wants to be able to pick them from a classification system.

  1. Describe the differences between supervised and unsupervised learning, including real-world examples of each.
  2. Preprocess data for unsupervised learning.
  3. Cluster data using the K-means algorithm.
  4. Determine the best number of centroids for K-means using the elbow curve.
  5. Use PCA to limit features and speed up the model.

Results:

The imported DataFrame before cleaning, 1252 rows of data:

Pic 1

The list of tradable cryptocurrencies after cleaning, 532 rows of data:

Pic 2

K-means Clustering Algorithm, Elbow Curve:

Pic 3

Applying the Principal Component Analysis:

Pic 4

3D Scatterplot with Clusters, Visualizing Tradable Cryptocurrencies:

Pic 5

Number of Tradable Cryptocurrencies:

Pic 6

DataFrame to plot results:

Pic 7

Tradable Cryptocurrencies:

Pic 8