next up previous contents
Next: Neighborhood function Up: Training Previous: Training

Update rule

By this update procedure described above, the net forms an elastic net that during learning folds onto the ``cloud'' formed by the input data. The codebook vectors tend to drift there where the data is dense, while there tends to be only a few codebook vectors where data is sparsely located. In this manner, the net tends to approximate the probability density function of the input data [20].

The Self-Organizing Map update rule for a unit tex2html_wrap_inline1965 is the following:

displaymath1967

where t denotes time. This is, as mentioned above, a training process through time. The tex2html_wrap_inline1971 is the input vector drawn from the input data set at time t. tex2html_wrap_inline1975 is a non-increasing neighborhood function around the winner unit tex2html_wrap_inline1943 . More on the subject of the neighborhood function in the next section.



Jaakko Hollmen
Fri Mar 8 13:44:32 EET 1996