Ignore:
Timestamp:
Feb 27, 2008, 5:58:15 PM (14 years ago)
Author:
Markus Ringnér
Message:

Working on #75

File:
1 edited

Legend:

Unmodified
Added
Removed
  • trunk/yat/classifier/SupervisedClassifier.h

    r1162 r1176  
    4343
    4444  ///
    45   /// @brief Interface class for supervised classifiers
     45  /// @brief Interface class for supervised classifiers that use data
     46  /// in a matrix format with data points as columns and each row
     47  /// corresponding to a variable for the data points. Supervised
     48  /// classifiers that do not use data in this format include
     49  /// kernel-based classifiers such as SVM.
    4650  ///
    47 
    4851  class SupervisedClassifier
    4952  {
     
    6366
    6467    ///
    65     /// An interface for making new classifier objects. This function
    66     /// allows for specification at run-time of which classifier to
    67     /// instatiate (see 'Prototype' in Design Patterns).
     68    /// An interface for making new %classifier objects. This function
     69    /// allows for specification at run-time of which %classifier type
     70    /// to instatiate (see 'Prototype' in Design Patterns). Derived
     71    /// classes should implement this function with DerivedClass* as
     72    /// the return type and not SupervisedClassifier*, and a
     73    /// dynamically allocated %classifier should be returned. The
     74    /// implementation of this function should correspond to a copy
     75    /// constructor with the exception that the returned %classifier
     76    /// is not trained.
    6877    ///
    69     /// @note Returns a dynamically allocated SupervisedClassifier, which has
     78    /// @note Returns a dynamically allocated %classifier, which has
    7079    /// to be deleted by the caller to avoid memory leaks.
    7180    ///
     
    8796
    8897    ///
    89     /// Train the classifier.
     98    /// Train the %classifier using unweighted training data and
     99    /// targets. The training data \a data should have one column per
     100    /// training sample and one row for each variable measured for the
     101    /// training samples. The size of \a target should be the number
     102    /// of samples in \a data and \a target should contain the class
     103    /// for each sample ordered in the same order as columns in \a data.
    90104    ///
    91     virtual void train(const MatrixLookup&, const Target&)=0;
     105    virtual void train(const MatrixLookup& data, const Target& targets)=0;
    92106
    93107    ///
    94     /// Train the classifier.
     108    /// Train the %classifier using weighted training data and
     109    /// targets. Both \a data and \a targets should follow the
     110    /// description for train(const MatrixLookup& data, const Target& targets)
    95111    ///
    96     virtual void train(const MatrixLookupWeighted&, const Target&)=0;
     112    virtual void train(const MatrixLookupWeighted& data, const Target& targets)=0;
    97113
    98114  }; 
Note: See TracChangeset for help on using the changeset viewer.