Self-Organizing Maps Extension

Introduction

The Self-Organizing Map (SOM) is a competitive neural network that provides a topological mapping from the input space to the clusters. One of its key features is the availability of powerful visualizations. It can be used

1. to get a visualized overview of multidimensional data,
2. to cluster data, and
3. to find non-linear relationships and dependencies between variables.

Using SOM in Davisor Chart

The SOM comes in SOM Extension. When it is installed into Davisor Chart there are new DataSources and ChartTypes available.

SOM algorithm is implemented as a processing data source. It takes one data source in and provides the Self-Organized Map out. More information about input data requirements, XML tags, and output data type are in Data Import Chapter.

When SOM is constructed it (or its components) can be visualized using the available visualization methods described with more details in Visualizations Chapter.

Typical use cases and instructions for using the existing maps are given in Chapter: Using Existing Maps.

For background information the SOM Theory is desribed in an own chapter. However, the SOM Option Pack can be used without exact knowledge of the algorithm.

Installing SOM Extension

SOM Extension is in use, when the davisorsom.jar is available in CLASSPATH. Typically this is done by copying the JAR file into your web application's WEB-INF/lib folder. Davisor Chart must also be installed into your web application.