Learning from ubiquitous data streams: Clustering data and data sources

摘要

Knowledge discovery techniques try to extract patterns and concepts from raw data, and clustering certainly is one of the most popular processes in this research field. However, nowadays data is being produced in streaming fashion and distributed locations, turning most classical methods obsolete. This thesis addresses two different clustering problems in ubiquitous and streaming scenarios, presenting evidence of the advantages produced by applying distributed and streaming machine learning algorithms, and proposing new ones to solve the addressed problems.