|Code Listing by Will Dwinnell|
Routine accepts input variables, binary target variable (0/1) and a small number of training parameters, and returns discovered coefficients for a single neuron.
Learning rule: incremental delta rule
Learning Rate: constant
Transfer function: logistic
Exemplar presentation order: random, by training epoch
Use 'help DeltaRule' for further details
Features of this implementation of LDA:
- Allows for >2 classes
- Permits user-specified prior probabilities
- Requires only base MATLAB (no toolboxes needed)
- Assumes that the data is complete (no missing values)
This routine calculates the median squared error of a linear function, and can be used with fminsearch as a robust linear regression (see help LinearMedianSquaredError).
It should be very easy to extend the example code to handle...
All files and free downloads are copyright of their respective owners. We do not provide any hacked, cracked, illegal, pirated version of scripts, codes, components downloads. All files are downloaded from the publishers website, our file servers or download mirrors. Always Virus check files downloaded from the web specially zip, rar, exe, trial, full versions etc. Download links from rapidshare, depositfiles, megaupload etc not published.