testinstances.java

Java 编写的多种数据挖掘算法 包括聚类、分类、预处理等

    
    if (Utils.getOptionPos("word-separators", options) > -1) {
      tmpStr = Utils.getOption("word-separators", options);
      setWordSeparators(tmpStr);
    }
    else if (!initialized) {
      setWordSeparators(DEFAULT_SEPARATORS);
    }
    
    tmpStr = Utils.getOption("date", options);
    if (tmpStr.length() != 0)
      setNumDate(Integer.parseInt(tmpStr));
    else if (!initialized)
      setNumDate(0);
    
    tmpStr = Utils.getOption("relational", options);
    if (tmpStr.length() != 0)
      setNumRelational(Integer.parseInt(tmpStr));
    else if (!initialized)
      setNumRelational(0);
    
    tmpStr = Utils.getOption("num-instances-relational", options);
    if (tmpStr.length() != 0)
      setNumInstancesRelational(Integer.parseInt(tmpStr));
    else if (!initialized)
      setNumInstancesRelational(10);
    
    if (!initialized)
      setMultiInstance(Utils.getFlag("multi-instance", options));
  }
  
  /**
   * Gets the current settings of this object.
   *
   * @return an array of strings suitable for passing to setOptions
   */
  public String[] getOptions() {
    Vector 	result;
    String[]	options;
    int		i;

    result = new Vector();
    
    result.add("-relation");
    result.add(getRelation());
    
    result.add("-seed");
    result.add("" + getSeed());
    
    result.add("-num-instances");
    result.add("" + getNumInstances());
    
    if (getNoClass()) {
      result.add("-no-class");
    }
    else {
      result.add("-class-type");
      result.add("" + getClassType());
      
      result.add("-class-values");
      result.add("" + getNumClasses());
      
      result.add("-class-index");
      result.add("" + getClassIndex());
    }
    
    result.add("-nominal");
    result.add("" + getNumNominal());
    
    result.add("-nominal-values");
    result.add("" + getNumNominalValues());
    
    result.add("-numeric");
    result.add("" + getNumNumeric());
    
    result.add("-string");
    result.add("" + getNumString());
    
    result.add("-words");
    result.add("" + getWords());
    
    result.add("-word-separators");
    result.add("" + getWordSeparators());
    
    result.add("-date");
    result.add("" + getNumDate());
    
    result.add("-relation");
    result.add("" + getNumRelational());
    
    result.add("-num-instances-relational");
    result.add("" + getNumInstancesRelational());

    if (getMultiInstance())
      result.add("-multi-instance");

    if (getHandler() != null) {
      result.add("-W");
      result.add(getHandler().getClass().getName());
      if (getHandler() instanceof OptionHandler) {
	result.add("--");
	options = ((OptionHandler) getHandler()).getOptions();
	for (i = 0; i < options.length; i++)
	  result.add(options[i]);
      }
    }
    
    return (String[]) result.toArray(new String[result.size()]);
  }
  
  /**
   * sets the name of the relation
   * 
   * @param value 	the name of the relation
   */
  public void setRelation(String value) {
    m_Relation = value;
  }
  
  /**
   * returns the current name of the relation
   * 
   * @return 		the name of the relation
   */
  public String getRelation() {
    return m_Relation;
  }
  
  /**
   * sets the seed value for the random number generator
   * 
   * @param value	the seed
   */
  public void setSeed(int value) {
    m_Seed   = value;
    m_Random = new Random(m_Seed);
  }
  
  /**
   * returns the current seed value
   * 
   * @return 		the seed value
   */
  public int getSeed() {
    return m_Seed;
  }
  
  /**
   * sets the number of instances to produce
   * 
   * @param value	the number of instances
   */
  public void setNumInstances(int value) {
    m_NumInstances = value;
  }
  
  /**
   * returns the current number of instances to produce
   * 
   * @return the number of instances
   */
  public int getNumInstances() {
    return m_NumInstances;
  }
  
  /**
   * sets the class attribute type
   * 
   * @param value	the class attribute type
   */
  public void setClassType(int value) {
    m_ClassType = value;
    m_RelationalClassFormat = null;
  }
  
  /**
   * returns the current class type
   * 
   * @return the class attribute type
   */
  public int getClassType() {
    return m_ClassType;
  }
  
  /**
   * sets the number of classes
   * 
   * @param value	the number of classes
   */
  public void setNumClasses(int value) {
    m_NumClasses = value;
  }
  
  /**
   * returns the current number of classes
   * 
   * @return the number of classes
   */
  public int getNumClasses() {
    return m_NumClasses;
  }
  
  /**
   * sets the class index (0-based)
   * 
   * @param value	the class index 
   * @see #CLASS_IS_LAST
   * @see #NO_CLASS
   */
  public void setClassIndex(int value) {
    m_ClassIndex = value;
  }
  
  /**
   * returns the current class index (0-based), -1 is last attribute
   * 
   * @return 		the class index
   * @see #CLASS_IS_LAST
   * @see #NO_CLASS
   */
  public int getClassIndex() {
    return m_ClassIndex;
  }
  
  /**
   * whether to have no class, e.g., for clusterers; otherwise the class
   * attribute index is set to last
   * 
   * @param value whether to have no class
   * @see #CLASS_IS_LAST
   * @see #NO_CLASS
   */
  public void setNoClass(boolean value) {
    if (value)
      setClassIndex(NO_CLASS);
    else
      setClassIndex(CLASS_IS_LAST);
  }
  
  /**
   * whether no class attribute is generated
   * 
   * @return 		true if no class attribute is generated
   */
  public boolean getNoClass() {
    return (getClassIndex() == NO_CLASS);
  }
  
  /**
   * sets the number of nominal attributes
   * 
   * @param value	the number of nominal attributes
   */
  public void setNumNominal(int value) {
    m_NumNominal = value;
  }
  
  /**
   * returns the current number of nominal attributes
   * 
   * @return 		the number of nominal attributes
   */
  public int getNumNominal() {
    return m_NumNominal;
  }
  
  /**
   * sets the number of values for nominal attributes
   * 
   * @param value	the number of values
   */
  public void setNumNominalValues(int value) {
    m_NumNominalValues = value;
  }
  
  /**
   * returns the current number of values for nominal attributes
   * 
   * @return 		the number of values
   */
  public int getNumNominalValues() {
    return m_NumNominalValues;
  }
  
  /**
   * sets the number of numeric attributes
   * 
   * @param value 	the number of numeric attributes
   */
  public void setNumNumeric(int value) {
    m_NumNumeric = value;
  }
  
  /**
   * returns the current number of numeric attributes
   * 
   * @return 		the number of numeric attributes
   */
  public int getNumNumeric() {
    return m_NumNumeric;
  }
  
  /**
   * sets the number of string attributes
   * 
   * @param value 	the number of string attributes
   */
  public void setNumString(int value) {
    m_NumString = value;
  }
  
  /**
   * returns the current number of string attributes
   * 
   * @return 		the number of string attributes
   */
  public int getNumString() {
    return m_NumString;
  }

  /**
   * turns the comma-separated list into an array
   * 
   * @param value	the list to process
   * @return		the list as array
   */
  protected static String[] listToArray(String value) {
    StringTokenizer	tok;
    Vector		list;
    
    list = new Vector();
    tok = new StringTokenizer(value, ",");
    while (tok.hasMoreTokens())
      list.add(tok.nextToken());
    
    return (String[]) list.toArray(new String[list.size()]);
  }
  
  /**
   * turns the array into a comma-separated list
   * 
   * @param value	the array to process
   * @return		the array as list
   */
  protected static String arrayToList(String[] value) {
    String	result;
    int		i;
    
    result = "";
    
    for (i = 0; i < value.length; i++) {
      if (i > 0)
	result += ",";
      result += value[i];
    }
    
    return result;
  }
  
  /**
   * Sets the comma-separated list of words to use for generating strings. The
   * list must contain at least 2 words, otherwise an exception will be thrown.
   * 
   * @param value			the list of words
   * @throws IllegalArgumentException	if not at least 2 words are provided
   */
  public void setWords(String value) {
    if (listToArray(value).length < 2)
      throw new IllegalArgumentException("At least 2 words must be provided!");
    
    m_Words = listToArray(value);
  }
  
  /**
   * returns the words used for assembling strings in a comma-separated list.
   * 
   * @return		the words as comma-separated list
   */
  public String getWords() {
    return arrayToList(m_Words);
  }

  /**
   * sets the word separators (chars) to use for assembling strings.
   * 
   * @param value	the characters to use as separators
   */
  public void setWordSeparators(String value) {
    m_WordSeparators = value;
  }
  
  /**
   * returns the word separators (chars) to use for assembling strings.
   * 
   * @return		the current separators
   */
  public String getWordSeparators() {
    return m_WordSeparators;
  }
  
  /**
   * sets the number of data attributes
   * 
   * @param value	the number of date attributes
   */
  public void setNumDate(int value) {
    m_NumDate = value;
  }
  
  /**
   * returns the current number of date attributes
   * 
   * @return		the number of date attributes
   */
  public int getNumDate() {
    return m_NumDate;
  }
  
  /**
   * sets the number of relational attributes
   * 
   * @param value	the number of relational attributes
   */
  public void setNumRelational(int value) {
    m_NumRelational = value;
    m_RelationalFormat = new Instances[value];
  }
  
  /**
   * returns the current number of relational attributes
   * 
   * @return		the number of relational attributes
   */
  public int getNumRelational() {
    return m_NumRelational;
  }
  
  /**
   * sets the number of instances in relational/bag attributes to produce
   * 
   * @param value	the number of instances
   */
  public void setNumInstancesRelational(int value) {
    m_NumInstancesRelational = value;
  }
  
  /**
   * returns the current number of instances in relational/bag attributes to produce
   * 
   * @return		the number of instances
   */
  public int getNumInstancesRelational() {
    return m_NumInstancesRelational;
  }

  /**
   * sets whether multi-instance data should be generated (with a fixed
   * data structure)
   * 
   * @param value	whether multi-instance data is generated
   */
  public void setMultiInstance(boolean value) {
    m_MultiInstance = value;
  }
  
  /**
   * Gets whether multi-instance data (with a fixed structure) is generated
   * 
   * @return		true if multi-instance data is generated
   */
  public boolean getMultiInstance() {
    return m_MultiInstance;
  }
  
  /**
   * sets the structure for the bags for the relational attribute
   * 
   * @param index       the index of the relational attribute
   * @param value       the new structure
   */
  public void setRelationalFormat(int index, Instances value) {
    if (value != null)
      m_RelationalFormat[index] = new Instances(value, 0);
    else
      m_RelationalFormat[index] = null;
  }
  
  /**
   * returns the format for the specified relational attribute, can be null
   * 
   * @param index       the index of the relational attribute
   * @return            the current structure
   */
  public Instances getRelationalFormat(int index) {
    return m_RelationalFormat[index];
  }

  /**
   * sets the structure for the relational class attribute
   * 
   * @param value	the structure for the relational attribute
   */
  public void setRelationalClassFormat(Instances value) {
    if (value != null)
      m_RelationalClassFormat = new Instances(value, 0);
    else
      m_RelationalClassFormat = null;
  }