MCQ IN COMPUTER SCIENCE & ENGINEERING

Detailed explanation-1: -Graph-based clustering performance can easily be improved by applying ICA blind source separation during the graph Laplacian embedding step. Applying unsupervised feature learning to input data using either RICA or SFT, improves clustering performance.

Detailed explanation-2: -The silhouette coefficient may provide a more objective means to determine the optimal number of clusters. This is done by simply calculating the silhouette coefficient over a range of k, and identifying the peak as the optimum K.

Detailed explanation-3: -The bigger number of the clusters will become harder to interpret the character of each cluster. However, the smaller number of the clusters obviously might not be able to capture a small but important difference between the groups that could have been found by increasing the number.

Detailed explanation-4: -What Is Good Clustering? The quality of a clustering result depends on both the similarity measure used by the method and its implementation. The quality of a clustering method is also measured by its ability to discover some or all of the hidden patterns.