Warning:
JavaScript is turned OFF. None of the links on this page will work until it is reactivated.
If you need help turning JavaScript On, click here.
This Concept Map, created with IHMC CmapTools, has information related to: Self Organizing Maps SOM, competitive Learning steps calculate distance between input and all weight vectors, Self-Organizing Map (SOM) has weight vectors, Self-Organizing Map (SOM) include Learning Algorithm, Self-Organizing Map (SOM) used in clustering, visualization, initial values can be regular array of vectorial values that lie on the subspace spanned by the eigenvectors corresponding to the two largest principal components of input data, Learning Algorithm has initial values, Self-Organizing Map (SOM) is Architecture, Self-Organizing Map (SOM) property topology is preserved, two modes of operation are mapping process, weight vectors have the same dimensionality as input vectors, θ(v,t) :Neighborhood Function can be 1 for neighbors 0 in other case, Self-Organizing Map (SOM) property Each input is connected to all output neurons, update formula is Wv(t+1)=Wv(t)+θ(v,t) α(t)(D(t)-Wv(t)), two modes of operation are training process, training uses competitive Learning, Self-Organizing Map (SOM) is single layer feedforward network, Wv(t+1)=Wv(t)+θ(v,t) α(t)(D(t)-Wv(t)) where Wv(t): weight vector of neuron v, Self-Organizing Map (SOM) is mapping from high dimensional spaces to 2D (or 3D) space, Self-Organizing Map (SOM) trained unsing Unsupervised Learning, stablish winner output called Best Matching Unit BMU