Changeset 4052


Ignore:
Timestamp:
Mar 26, 2021, 3:26:19 AM (7 months ago)
Author:
Peter
Message:

add description on what is optimised during the training.

Location:
trunk/yat/classifier
Files:
2 edited

Legend:

Unmodified
Added
Removed
  • trunk/yat/classifier/Perceptron.cc

    r3709 r4052  
    107107        y(i) = 1.0;
    108108
     109    // We use the Iteratively Rewighted Least Square algorithm as described
     110    // https://en.wikipedia.org/wiki/Logistic_regression
     111
    109112    size_t max_epochs = 100;
    110113    double sum_squared = 1.0; // some (relatively) large number
  • trunk/yat/classifier/Perceptron.h

    r3709 r4052  
    3535     \brief A Single-layer Perceptron
    3636
    37      Data are modeled as \f$ y = \frac{1}{1 + \exp(-wx)} \f$
     37     Data are modeled as \f$ \mu = \frac{1}{1 + \exp(-wx)} \f$
    3838
    3939     \since New in yat 0.16
     
    8686       \brief train the model
    8787
     88       Model parameters, \f$w\f$, are calculated such that the
     89       log-likelihood,
     90       \f$ \log \mathcal{L} = \sum y_i \log \left(\mu_i\right) +
     91       (1-y_i) \log \left(1 - \mu_i\right) \f$,
     92       is maximized.
     93
    8894       \param x each row corresponds to a data point and each column a
    8995       feature.
Note: See TracChangeset for help on using the changeset viewer.