What is the average Silhouette score of your final k-means model (enter with 4 decimal digits)?

Load the Mall Customers data set. Study the plot and the table view to understand what is shown on the plot. Then apply a k-means clustering model and take a look at the inertia elbow plot (attach the plot here). Decide on a suitable number of clusters and create the final model. Also attach the cluster plot of your final model (that also shows the centroids and the Voronoi tesselation). Investigate the clusters: give a brief description of each one, how can we interpret them (which kind of customer do they represent)?

What is the optimal number of clusters in the Mall Customers dataset?

Apart from k-means, which other methods work on the Penguins data set and deliver a good separation of the 3 groups?

Comment on these values. In general, what are good/bad silhouette and inertia values? Are they in this case here how you would expect them to be - do you note any surprising values? If yes/no, why could that be? 

Attach the two silhouette plots for k=2 and k=3.

Investigate the clustering result and compare it with the plot of the true groups. Find a sample that has been classified incorrectly. Take a look at the sample on the plot, report its index and the assigned probabilities here. Why do you think was the classification done incorrectly by the model?

Load the concentric circles data with the following parameters:

• N: 500
• Noise: 0.1
• Inner circle size: 0.35

Decide on a cluster method (there is one correct choice) and create a model for the concentric circles data. Optimize your hyperparameter(s) so that you get a good representation of the two clusters.

In your answer, report the method you used and why, your hyperparameter settings for your method and include a scatterplot of the clustering.

Find one unexpectedly close word pair in the 2D view. Report the pair and their cosine similarity. Then explain what this implies about 2D projections and give a plausible reason for the apparent proximity.