To figure out how many labels the clustering finds and how many members each label has. Also for visualization, perform a 2 PCA transformation and use PC1 and PC2 as the X and Y axes of a 2-D scatter plot to show all data points. Show the different DBSCAN clusters in different colors on the data points.
Things to do with the PCA results:
Perform a PCA analysis with number of components from 2 to 300
Show a plot where on the X-axis the number of components is shown and on the Y-Axis the total explained variance using this number of components.
5 freelancers estão ofertando em média ₹1980 nesse trabalho
Hi, I'm an ml engineer with a 2y experience at CV and NLP. I have experience at similar tasks, like w2v/BERT visualization and clustering, and finding an optimal number of clusters.