We are interested in practical tools for the quantitative evaluation of self-organizing maps (SOMs). Recently it has been argued that any quality measure for SOMs needs to evaluate the embedding or coverage of a map as well as its topological quality. Over the years many different quality measures for self-organizing maps have been proposed. However, many of these only measure one aspect of a S...

We propose a way of creating product maps with self-organizing maps (SOMs) for purchase decision making. We previously proposed a way of purchase decision support using SOMs and the Analytic Hierarchy Process (AHP). We provided several class boundaries, which divided the input features into several classes before creating self-organizing product maps. Because the number of classes and their bou...

Title of Dissertation: One-Shot Multi-Winner Self-Organizing Maps Reiner Schulz, Doctor of Philosophy, July 22, 2004 Dissertation directed by: Dr. James Reggia Department of Computer Science There exist two different approaches to self-organizing maps (SOMs). One approach, rooted in theoretical neuroscience, uses SOMs as computational models of biological cortex. The other approach, taken in co...

This paper proposes an extension of an SOM called the “SOM of SOMs,” or SOM, in which objects to be mapped are self-organizing maps. In SOM, each nodal unit of a conventional SOM is replaced by a function module of SOM. Therefore, SOM can be regarded as a variation of a modular network SOM (mnSOM). Since each child SOM module in SOM is trained to represent an individual map, the parent map in S...

Self-Organizing Maps, or Kohonen networks, are a widely used neural network architecture. This paper starts with a brief overview of how selforganizing maps can be used in different types of problems. A simple and intuitive explanation of how a self-organizing map is trained is provided, together with a formal explanation of the algorithm, and some of the more important parameters are discussed...

In this study, we visualize Pareto-optimum solutions derived from multiple-objective optimization using spherical self-organizing maps (SOMs) that lay out SOM data in three dimensions. There have been a wide range of studies involving plane SOMs where Pareto-optimal solutions are mapped to a plane. However, plane SOMs have an issue that similar data differing in a few specific variables are oft...

This paper presents a methodology for missing value imputation. The methodology is based on a combination of Self-Organizing Maps (SOM), where combination is achieved by Nonnegative Least Squares algorithm. Instead of a need for validation as when using traditional SOMs, the combination proceeds straight into final model building. Therefore, the methodology has very low computational time. The ...

Overfitting in multilayer perceptron (MLP) training is a serious problem. The purpose of this study is to avoid overfitting in on-line learning. To overcome the overfitting problem, we have investigated feeling-of-knowing (FOK) using self-organizing maps (SOMs). We propose MLPs with FOK using the SOMs method to overcome the overfitting problem. In this method, the learning process advances acco...

Self organizing maps (SOMs) are widely-used for unsupervised classification. For this application, they must be combined with some partitioning scheme that can identify boundaries between distinct regions in the maps they produce. We discuss a novel partitioning scheme for SOMs based on the Bayesian Blocks segmentation algorithm of Scargle [1998]. This algorithm minimizes a cost function to ide...

In a recent paper, T. Kohonen and P. Somervuo have shown that self-organizing maps (SOMs) are not restricted to numerical data. They can also be defined for symbol strings, provided that one defines an average function for strings and that the adaptation process is performed off-line (batch). In this paper, we present two different methods for computing averages of strings, as well as an on-lin...