Climbing Robots

摘要

Based on the Self-Organizing Map (SOM) algorithm, development of effective solutions for visual analysis and retrieval in complex data is possible. Example application domains include retrieval in multimedia data bases, and analysis in financial, text, and general highdimensional data sets. While early work defined basic concepts for data representation and visual mappings for SOM-based analysis, recent work contributed advanced visual representations of the output of the SOM algorithm, and explored innovative application concepts. In this article, we review a selection of classic and more recent approaches to SOM-based visual analysis. We argue that important improvements have been achieved which allow effective visual representation and interaction with the output of the SOM algorithm. We identify promising directions for future research, which will support new application areas and provide additional advanced visualization approaches.