Self-Organizing Maps (Springer Series in Information Sciences)

Springer, 1997 8vo (23.5 cm), XVII, 426 pp. Printed wrappers. A seminal work that unveils the groundbreaking concept of self-organizing maps (SOMs) and their profound impact on the field of artificial intelligence. In this influential book, Kohonen introduces an innovative neural network architecture that mimics the self-organizing principles of the human brain, enabling machines to efficiently analyze and classify complex data patterns. Kohonen's work revolutionizes the field of unsupervised learning by presenting a powerful algorithmic approach that enables machines to learn and adapt without explicit guidance. The self-organizing map algorithm constructs a low-dimensional representation of high-dimensional data, effectively capturing the inherent structure and relationships within the input space. By arranging data samples on a map-like grid, SOMs reveal hidden patterns, clusters, and topological relationships, facilitating data exploration and visualization. The book provides a comprehensive introduction to SOMs, offering a clear and accessible explanation of the underlying principles, mathematical foundations, and algorithmic implementation. Kohonen's intuitive and lucid writing style guides readers through the intricate details of SOMs, making the concepts approachable for both novices and experts in the field. Moreover, "Self-Organizing Maps" showcases the broad range of applications where SOMs have proven invaluable. From data mining and pattern recognition to image and speech processing, Kohonen demonstrates the versatility of SOMs in various domains. The book illustrates real-world case studies and practical examples, emphasizing the potential of SOMs as a powerful tool for solving complex problems across multiple disciplines.