Temporal Data Mining via Unsupervised Ensemble Learning
Par :Formats :
- Réservation en ligne avec paiement en magasin :
- Indisponible pour réserver et payer en magasin
- Nombre de pages157
- PrésentationBroché
- FormatGrand Format
- Poids0.38 kg
- Dimensions19,0 cm × 23,5 cm × 0,8 cm
- ISBN978-0-12-811654-8
- EAN9780128116548
- Date de parution01/01/2017
- ÉditeurElsevier
Résumé
Key Features : The first novel approach is based on the ensemble of Hidden Markov Model-based partitioning clustering, associated with a hierarchical clustering refinement, to solve problems by finding the intrinsic number of clusters and model initialization problems which exist in most model-based clustering algorithms ; The second approach presents an unsupervised ensemble learning model of iteratively constructed partitions on a sub-training set obtained by a hybrid sampling scheme which provides a potential solution for large temporal data clustering tasks ; The third proposed approach is a feature-based approach to temporal data clustering, through a weighted ensemble of a simple clustering algorithm with minimum user-dependent parameters, to address both proper grouping with minimum computational cost and provide a generic technique for the optimal solution of combining multiple partitions.
Temporal Data Mining via Unsupervised Ensemble Learning not only enumerates the existing techniques proposed so far, but also classifies and organizes them in a way that is of help for a practitioner looking for solutions to a concrete problem. The evidence suggests that ensemble learning techniques may give an optimal solution for dealing with temporal data clustering problems, and this book presents the case in an accessible format designed to appeal to both students and professional researchers, including those with little mathematical and statistical background.