The beginning of the age of artificial intelligence and machine learning has created new challenges and opportunities for data analysts, statisticians, mathematicians, econometricians, computer scientists and many others. At the root of these techniques are algorithms and methods for clustering and classifying different types of large datasets, including time series data. Time Series Clustering and Classification includes relevant developments on observation-based, feature-based and model-based traditional and fuzzy clustering methods, feature-based and model-based classification methods, and machine learning methods. It presents a broad and self-contained overview of techniques for both researchers and students. Features Provides an overview of the methods and applications of pattern recognition of time series Covers a wide range of techniques, including unsupervised and supervised approaches Includes a range of real examples from medicine, finance, environmental science, and more R and MATLAB code, and relevant data sets are available on a supplementary website

Classification and clustering of time series is becoming an important area of research in several fields, such as economics, marketing, business, finance, medicine, biology, physics, psychology, zoology, and many others. For example, in economics we may be interested in classifying the economic situation of a country by looking at some time series indicators, such as Gross National Product, disposable income, unemployment rate or inflation rate. In this book, we propose new measures of distance between time series based on the autocorrelations, partial and inverse autocorrelations, and periodogram ordinates. The use of both hierarchical and nonhierarchical clustering algorithms is considered. We also introduce time and frequency domain based metrics for classification of time series with unequal lengths. As economic applications, we present two illustrative examples. The first uses economic time series data to identify similarities among industrial production series in the United States. The second applies the interpolated periodogram based method for classifying time series with unequal lengths of industrialized countries.