Some Machine Learning Approaches to the Analysis of Temporal Data
share
Investigating time is not restricted to time series analysis, where from a sequence of equidistant measurements the value of the next measurement is predicted. In contrast, many applications have to cope with very large collections of time series data. The tasks range from regression and classification to detecting patterns in the data. By several case studies stemming from several years of research, this chapter illustrates the diversity of temporal phenomena handled in machine learning and data mining on the basis of very large data sets. The path leads from time series classification to the analysis of streaming data. A recurrent theme is the appropriate representation, feature extraction, and feature selection for high performance learning.
READ FULL TEXT
