دانلود کتاب جنبه های نظری و عملی خوشه بندی

Clustering is a part of machine learning that seeks to identify groups into sets of objects such that objects that belong to the same group are as similar as possible, and objects that belong to two distinct groups are as dissimilar as possible. In general, clustering exploration is based on computing similarities (or dissimilarities) between objects but does not provide the reasons for the existence of these groupings. Various notions of dissimilarities are considered among objects ranging from simple dissimilarities, metrics on linear spaces, ultrametrics, and extensions of these measures to sets. Studying these measures requires incursions in a variety of mathematical disciplines ranging from linear algebra and optimization to functional analysis and topology. The results of clusterings are evaluated using a variety of criteria allowing users to choose clusterings that are desirable from the point of view of these criteria. Clustering use is widespread, ranging from genomics, epidemiology, medicine, economics and many other disciplines. The intended readership of this volume consists of researchers and graduate students who work in data mining and pattern recognition, or apply those in their domain of interest. I strived to make this volume as self-contained as possible. Appendices, exercises, and supplements are provided to help readers in their search of mathematical tools useful for clustering.