Unsupervised Learning

Beschreibung

A new approach to unsupervised learning Evolving technologies have brought about an explosion ofinformation in recent years, but the question of how suchinformation might be effectively harvested, archived, and analyzedremains a monumental challenge--for the processing of suchinformation is often fraught with the need for conceptualinterpretation: a relatively simple task for humans, yet an arduousone for computers. Inspired by the relative success of existing popular research onself-organizing neural networks for data clustering and featureextraction, Unsupervised Learning: A Dynamic Approachpresents information within the family of generative,self-organizing maps, such as the self-organizing tree map (SOTM)and the more advanced self-organizing hierarchical variance map(SOHVM). It covers a series of pertinent, real-world applicationswith regard to the processing of multimedia data--from itsrole in generic image processing techniques, such as the automatedmodeling and removal of impulse noise in digital images, toproblems in digital asset management and its various roles infeature extraction, visual enhancement, segmentation, and analysisof microbiological image data. Self-organization concepts and applications discussedinclude: * Distance metrics for unsupervised clustering * Synaptic self-amplification and competition * Image retrieval * Impulse noise removal * Microbiological image analysis Unsupervised Learning: A Dynamic Approach introduces anew family of unsupervised algorithms that have a basis inself-organization, making it an invaluable resource forresearchers, engineers, and scientists who want to create systemsthat effectively model oppressive volumes of data with little or nouser intervention.

Autor*in

Matthew Kyan

Themen in »Unsupervised Learning«

Computer Science Computer Science Special Topics Data Mining Data Mining Statistics Electrical & Electronics Engineering Elektrotechnik u. Elektronik Informatik Intelligent Systems & Agents Intelligente Systeme u. Agenten Spezialthemen Informatik Statistics Statistik

Stimmen zu »Unsupervised Learning«

Details

Link teilen