This video explains three different unsupervised clustering algorithms: k-means clustering, spectral clustering, and DBSCAN (Density-Based Spatial Clustering of Applications with Noise). Clustering algorithms are essential unsupervised learning techniques that aim to find groups of similar instances in an unlabeled data set. We discuss different cost functions and approximation methods to solve them, including scikit-learn implementations with important input parameters. We also show the connections between k-means clustering and spectral clustering. The spectral clustering algorithm forms an affinity graph and the normalized Laplacian matrix is constructed. The eigenvalue decomposition of the Laplacian matrix is used to find a matrix of size n by k, followed by applying the k-means clustering algorithm to the rows of the resulting matrix. We use several synthetic or simulated data sets to exhibit the performance of different clustering methods using Google colab.
Link to the SVD video: [ Ссылка ]
#Clustering #UnsupervisedLearning #Spectralclustering
