Here are some of the algorithms for machine learning mostly clustering.....
The main topics included are:
- K means clustering
- Image clustering using K means:
Kmeans used to cluster images
- K means using Gradient Descent:
As in the paper "Convergence properties of kmeans" by Leon Bottou and Yoshua Bengio k means can be expressed as a gradient descent algorithm with superlinear convergence rate (It's equivalent to Newton's method). This is the implementation of kmeans as gradient descent.
- Graphs of Cost-Function versus centroids:
It includes the graph of cost function of k means for one dimensional data (euclidean distance) and shows that for one dimension the cost function in concave upward and hence with unique local minimum.
- Min-Max Kmeans algorithm:
This algorithm aims to solve the initialisation problems in k-means by minimising the size(variance) of the largest cluster.It effectively tries to form clusters of more or less same sizes.