Competitive Learning and Self‐Organizing Map
Kelvin K. L. Wong · 2023
The self-organizing map is one of the most widely used unsupervised artificial neural networks in which the system has no prior knowledge of the input data's properties or qualities, as well as the output data's class labels. Competitive learning is a type of Artificial Neural Network learning in which various neurons or processing units compete to learn how to represent current input. A Growing Self-Organizing Map begins with only two neurons, resulting in a one-dimensional arrangement. The time adaptive self-organizing map network is a modified self-organizing map network with adaptive learning rates and neighborhood sizes as its learning parameters. A self-organizing map can be turned into an Oriented and Scalable Map by generalizing the neighborhood function and the winner selection. The Generative Topographic Mapping model is a probability density model that depicts the distribution of data in a multi-dimensional space using a fewer number of latent (or hidden) variables.