The document discusses unsupervised learning, focusing on clustering as a method to group similar instances without labeled data. It details the k-means algorithm for efficient clustering and highlights improvements like k-means++ for better centroid initialization. Additionally, it covers essential k-means functionalities and parameter settings for optimal performance.

1. Unsupervised Learning
DR. MAJID ALI KHAN

2. Unsupervised Learning
- Learning in the absence of labelled data
◦ Dimensionality Reduction
◦ Clustering
◦ Anamoly Detection

3. Clustering
Clustering: It is the task of identifying similar instances and assigning
them to clusters, i.e. create groups of similar instances.
Just like in classification, each instance gets assigned to a group.
However, unlike classification, clustering is an unsupervised task.

4. Clustering vs. Classification

5. K-Means
Clustering
The K-Means algorithm is a simple algorithm capable of clustering this kind of
dataset very quickly and efficiently, often in just a few iterations. It was proposed
by Stuart Lloyd at Bell Labs in 1957

6. K-Means in Scikit

7. K-Means Decision Boundaries

8. How Kmeans
works?

9. KMeans can give suboptimal solutions!!!

10. K-Means++
K-Means++ introduced a smarter initialization step that tends to select
centroids that are distant from one another, and this improvement makes
the K-Means algorithm much less likely to converge to a suboptimal solution.
- Default implementation of K-Means implements K-Means++ unless you
override the default init parameters with init=“random”
- Default implementation also uses n_init = 10, which means it’s going to run
K-Means ten times with different random initializations and choose the best
model

11. Choosing the right k?

12. Kmeans
kmeans.labels_
kmeans.cluster_centers_
kmeans.predict(X_new)
kmeans.inertia_
kmeans.score(X)