Spherical Self-Organizing Map
Budget: $250 – $750 CAD
Build Spherical Self-Organizing Map to learn human mouse movement data in order to confirm if the neurons in the border of the previous map were affected by the “border effect” of SOM or not.
In the latest results we received from SOM upon learning it with human mouse data, we saw that most of the neurons were placed on the border of the map, this problem is called, “border effect” of SOM, in which border neurons in the SOM map do not ‘stretch out’ during the training process as much as they should, and as a result they tend to ‘attract’ many potentially very different/distant points located on the ‘outside’ of the SOM border. This phenomenon in known in the literature as ‘SOM border effect’.
* To avoid “border effect” of SOM we can use Spherical Self-Organizing Maps
In the latest results we received from SOM upon learning it with human mouse data, we saw that most of the neurons were placed on the border of the map, this problem is called, “border effect” of SOM, in which border neurons in the SOM map do not ‘stretch out’ during the training process as much as they should, and as a result they tend to ‘attract’ many potentially very different/distant points located on the ‘outside’ of the SOM border. This phenomenon in known in the literature as ‘SOM border effect’.
* To avoid “border effect” of SOM we can use Spherical Self-Organizing Maps