001 /**
002 *
003 */
004 package org.wdssii.decisiontree;
005
006 import java.beans.XMLDecoder;
007 import java.io.BufferedReader;
008 import java.io.File;
009 import java.io.FileInputStream;
010 import java.io.FileReader;
011 import java.io.FileWriter;
012 import java.io.PrintWriter;
013 import java.util.ArrayList;
014 import java.util.List;
015
016 /**
017 *
018 * Uses a trained decision tree to classify cases. If labels are provided, also reports skill score.
019 *
020 * @author lakshman
021 *
022 */
023 public class Classify {
024 /**
025 * Classifies all inputs in a file based on a trained decision tree
026 */
027 public static void main(String[] args) throws Exception {
028 if (args.length < 2) {
029 System.err
030 .println("Usage: java org.jscience.statistics.decisiontree.Classify decisiontree.xml testinfile.csv outdir [separator]\n");
031 System.err
032 .println("The data file should have each example on a line, with all the attributes being numeric and separated by commas.");
033 System.err
034 .println("The final attribute can be the true classification if known. In this case, skill will be computed.");
035 return;
036 }
037
038 String decisiontree = args[0];
039 String infile = args[1];
040 String outdir = args[2];
041 String separator = (args.length > 3) ? args[3] : ",";
042
043 XMLDecoder decoder = new XMLDecoder(new FileInputStream(decisiontree));
044 AxialDecisionTree tree = (AxialDecisionTree) decoder.readObject();
045
046 BufferedReader reader = null;
047 try {
048 new File(outdir).mkdirs();
049 // read the file
050 reader = new BufferedReader(new FileReader(infile));
051 String line;
052 List<String[]> testingData = new ArrayList<String[]>();
053 while ((line = reader.readLine()) != null) {
054 String[] splitValues = line.split(separator);
055 testingData.add(splitValues);
056 }
057 // get the data and categories
058 int numTesting = testingData.size();
059 if (numTesting == 0){
060 System.err.println("Empty file: not processed");
061 return;
062 }
063 int numAttrInFile = testingData.get(0).length;
064 int numAttr = tree.getNumAttributes();
065 if (numAttrInFile < numAttr || numAttrInFile > (numAttr+1) ){
066 // numAttr and numAttr+1 are ok
067 throw new IllegalArgumentException("The file contains only " + numAttrInFile + " columns but the decision tree was trained with " + numAttr);
068 }
069 float[][] data = new float[numTesting][numAttr];
070 int[] categories = (numAttrInFile > numAttr)? new int[numTesting] : null;
071 for (int i = 0; i < numTesting; ++i) {
072 String[] indata = testingData.get(i);
073 if (indata.length != numAttrInFile) {
074 throw new IllegalArgumentException("Row number " + (i + 1)
075 + " of file has " + indata.length
076 + " attributes. Expected " + numAttrInFile);
077 }
078 for (int j = 0; j < numAttr; ++j) {
079 try {
080 data[i][j] = Float.parseFloat(indata[j]);
081 } catch (NumberFormatException e) {
082 System.err.println(e.getMessage() + " in attribute no."
083 + (j + 1) + " of line " + (i + 1)
084 + " Assuming zero.");
085 data[i][j] = 0;
086 }
087 }
088 if ( categories != null ){
089 categories[i] = Integer.parseInt(indata[numAttr]);
090 }
091 }
092
093 // classify and compute skill
094 MulticategorySkillScore tss = new MulticategorySkillScore(tree.getNumCategories());
095 PrintWriter writer = new PrintWriter(new FileWriter(outdir+"/classified.txt"));
096 for (int i=0; i < numTesting; ++i){
097 int result = tree.classify(data[i]);
098 writer.println(result);
099 if (categories != null){
100 tss.update(categories[i], result);
101 }
102 }
103
104 if ( categories != null ){
105 System.out.println(tss);
106 PrintWriter skillWriter = new PrintWriter(new FileWriter(outdir
107 + "/testing_skill.txt"));
108 skillWriter.println(tss);
109 skillWriter.close();
110 }
111
112 } finally {
113 if (reader != null) {
114 reader.close();
115 }
116 }
117 }
118 }