You no longer need to be an expert on SOM algorithms to create, visualize, and analyze your own Self-Organizing Maps thanks to SOM Analyzer. If you happen to be an expert, you will find in this excellent tool all the elements that you need to make the most of SOMs and to visualize in two dimensions problems of high-dimensional data. If you are not, the program’s wizard will guide you through the entire SOM creation process in a straightforward and clear way.
Thus, the program offers you the possibility of creating SOMs in either full automatic mode or in semi-automatic mode or in full manual mode, according to your knowledge of how the SOM algorithm works. You can let the program deal with the SOM parameters and feed it with the data you wish to analyze, or you can be in full control of the creation process from the start by entering your own parameters. SOM Analyzer will then create the SOM following your preferences, and will let you train and test it against your own data sets. Both online and offline testing are supported so that you can feed your SOM with real-time data coming from an external source. Data can come in file format or directly from a MySQL database, to which you can connect the app via the MySQL Connector.
To help you create the SOM that better fits your data and your preferences, the program includes Iterator, a tool that will create various SOMs following your parameters so that you can choose the one that’s best for your analysis. The various visualization methods provided will also allow you to work with the one that fits your map best. Neurons can be labeled, thus adding meaning to your maps and helping others to understand them better. Besides, the high level of flexibility that SOM Analyzer offers will let you configure the program according to your preferred color schemes and working environment, and save those settings for future use. The possibilities are manifold, and they’re there for you to discover.
SOM Analyzer is a high-end tool to produce high-quality Self-Organizing Maps, to train and test them with your own data sets, and to discover correlations and properties that cannot be visualized otherwise.