Practical Machine Learning Cookbook

Self-organizing map - visualizing of heatmaps

Over the past decade, there has been exponential growth in information. Gaining new knowledge from such databases is difficult, costly, and time-consuming if done manually. It may even be impossible when the data exceeds certain limits of size and complexity. As a result, the automated analysis and visualization of massive multidimensional datasets have been the focus of much scientific research over the last few years. The principal objective of this analysis and visualization is to find regularities and relationships in the data, thereby gaining access to hidden and potentially useful knowledge. A self-organizing map (SOM) is an unsupervised neural network algorithm that projects high-dimensional ...