Review of "Introduction to clustering large and high-dimensional data" by J. Kogan

Roughly speaking, clustering is a data analysis task to group a set of items into different categories so that items within one category are similar and items between different categories are dissimilar, where similar and dissimilar depend on the definition of distance between items. Although known for many decades, recently clustering has gained a lot of importance due to the exponential growth of digital libraries and the World Wide Web and the thus resulting need to find and extract information. Motivated by these Information Retrieval (IR) applications, which are usually characterized by large, sparse and high-dimensional data, “Introduction to Clustering Large and High-Dimensional Data” by J. Kogan is a textbook that tries to focus on a few clustering techniques that are very common in IR. In particular, it focuses on the kmeans algorithm, which is by far the most popular one in IR, including many of its variations, among them incremental kmeans, spherical k-means, quadratic k-means, k-means with divergences and others.