Icon

Solution 6 Training Clustering Algorithms

This workflow shows a solution to a hands-on exercise in the L4-ML Introduction to Machine Learning Algorithms self-paced course

Task 1: Perform k-Means clustering1. Create 3 clusters using the k-Means clustering algorithm2. Calculate the overall mean silhouette coefficient3. Change the number of clusters to 4 and calculate the mean overall silhouettecoefficient again Task 2: Compare hierarchical and k-Means clustering1. Perform hierarchical clustering on the data and select 4 clusters2. Assign colors to the clusters produced by the k-Means algorithm (from task1) and the hierarchical clustering algorithm3. Compare the clusters visually via a scatter plot for each algorithm Task 3: Compare the results between different configurations of the DBSCANalgorithm1. Calculate the Euclidean distances between the features 0 and 12. Perform DBSCAN clustering using the default configuration 3. Visualize the clusters produced by the DBSCAN in a scatter plot4. Change the Epsilon value to 0.55. Keep Epsilon value at 0.5 and increase the Minimum points to 5 k-Meanshierarchical clusteringEuclideanReadclustering-data.table HierarchicalClustering and Heatmap k-Means Color Manager Scatter Plot SilhouetteCoefficient DBSCAN Color Manager Scatter Plot Numeric Distances Color Manager Scatter Plot Table Reader Task 1: Perform k-Means clustering1. Create 3 clusters using the k-Means clustering algorithm2. Calculate the overall mean silhouette coefficient3. Change the number of clusters to 4 and calculate the mean overall silhouettecoefficient again Task 2: Compare hierarchical and k-Means clustering1. Perform hierarchical clustering on the data and select 4 clusters2. Assign colors to the clusters produced by the k-Means algorithm (from task1) and the hierarchical clustering algorithm3. Compare the clusters visually via a scatter plot for each algorithm Task 3: Compare the results between different configurations of the DBSCANalgorithm1. Calculate the Euclidean distances between the features 0 and 12. Perform DBSCAN clustering using the default configuration 3. Visualize the clusters produced by the DBSCAN in a scatter plot4. Change the Epsilon value to 0.55. Keep Epsilon value at 0.5 and increase the Minimum points to 5 k-Meanshierarchical clusteringEuclideanReadclustering-data.table HierarchicalClustering and Heatmap k-Means Color Manager Scatter Plot SilhouetteCoefficient DBSCAN Color Manager Scatter Plot Numeric Distances Color Manager Scatter Plot Table Reader

Nodes

Extensions

Links