Package ciir.umass.edu.learning

Class DataPoint

  • Direct Known Subclasses:
    DenseDataPoint, SparseDataPoint
    public abstract class DataPoint
extends java.lang.Object
    Author:
    vdang This class implements objects to be ranked. In the context of Information retrieval, each instance is a query-url pair represented by a n-dimentional feature vector. It should be general enough for other ranking applications as well (not limited to just IR I hope).

    • Constructor Summary

      Constructors 
      Modifier Constructor Description
      protected DataPoint()
      Default constructor.
      protected DataPoint​(java.lang.String text)
      The input must have the form:

    • Method Summary

      All Methods Static Methods Instance Methods Abstract Methods Concrete Methods 
      Modifier and Type Method Description
      double getCached()  
      java.lang.String getDescription()  
      abstract float getFeatureValue​(int fid)
      Get the value of the feature with the given feature ID
      abstract float[] getFeatureVector()
      Gets the value of all features as a dense array of feature values.
      java.lang.String getID()  
      protected static java.lang.String getKey​(java.lang.String pair)  
      float getLabel()  
      int getNumberOfKnownFeatures()  
      protected static java.lang.String getValue​(java.lang.String pair)  
      protected static boolean isUnknown​(float fVal)  
      protected float[] parse​(java.lang.String text)
      Parse the given line of text to construct a dense array of feature values and reset metadata.
      void resetCached()  
      void setCached​(double c)  
      void setDescription​(java.lang.String description)  
      abstract void setFeatureValue​(int fid, float fval)
      Set the value of the feature with the given feature ID
      abstract void setFeatureVector​(float[] dfVals)
      Sets the value of all features with the provided dense array of feature values
      void setID​(java.lang.String id)  
      void setLabel​(float label)  
      java.lang.String toString()  

      • Methods inherited from class java.lang.Object

        clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

    • Field Detail

      • missingZero

        public static boolean missingZero

      • MAX_FEATURE

        public static int MAX_FEATURE

      • FEATURE_INCREASE

        public static int FEATURE_INCREASE

      • UNKNOWN

        protected static float UNKNOWN

      • label

        protected float label

      • id

        protected java.lang.String id

      • description

        protected java.lang.String description

      • fVals

        protected float[] fVals

      • knownFeatures

        protected int knownFeatures

      • cached

        protected double cached

    • Constructor Detail

      • DataPoint

        protected DataPoint()
        Default constructor. No-op.

      • DataPoint

        protected DataPoint​(java.lang.String text)
        The input must have the form:
        Parameters:
        text -

    • Method Detail

      • isUnknown

        protected static boolean isUnknown​(float fVal)

      • getKey

        protected static java.lang.String getKey​(java.lang.String pair)

      • getValue

        protected static java.lang.String getValue​(java.lang.String pair)

      • parse

        protected float[] parse​(java.lang.String text)
        Parse the given line of text to construct a dense array of feature values and reset metadata.
        Parameters:
        text -
        Returns:
        Dense array of feature values

      • getFeatureValue

        public abstract float getFeatureValue​(int fid)
        Get the value of the feature with the given feature ID
        Parameters:
        fid -
        Returns:

      • setFeatureValue

        public abstract void setFeatureValue​(int fid,
                                     float fval)
        Set the value of the feature with the given feature ID
        Parameters:
        fid -
        fval -

      • setFeatureVector

        public abstract void setFeatureVector​(float[] dfVals)
        Sets the value of all features with the provided dense array of feature values

      • getFeatureVector

        public abstract float[] getFeatureVector()
        Gets the value of all features as a dense array of feature values.

      • getID

        public java.lang.String getID()

      • setID

        public void setID​(java.lang.String id)

      • getLabel

        public float getLabel()

      • setLabel

        public void setLabel​(float label)

      • getDescription

        public java.lang.String getDescription()

      • setDescription

        public void setDescription​(java.lang.String description)

      • setCached

        public void setCached​(double c)

      • getCached

        public double getCached()

      • resetCached

        public void resetCached()

      • toString

        public java.lang.String toString()
        Overrides:
        toString in class java.lang.Object

      • getNumberOfKnownFeatures

        public int getNumberOfKnownFeatures()