1 /* 2 * Copyright (C) 2008-2009 The Android Open Source Project 3 * 4 * Licensed under the Apache License, Version 2.0 (the "License"); 5 * you may not use this file except in compliance with the License. 6 * You may obtain a copy of the License at 7 * 8 * http://www.apache.org/licenses/LICENSE-2.0 9 * 10 * Unless required by applicable law or agreed to in writing, software 11 * distributed under the License is distributed on an "AS IS" BASIS, 12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. 13 * See the License for the specific language governing permissions and 14 * limitations under the License. 15 */ 16 17 package android.gesture; 18 19 import java.util.ArrayList; 20 import java.util.Collections; 21 import java.util.Comparator; 22 import java.util.TreeMap; 23 24 /** 25 * An implementation of an instance-based learner 26 */ 27 28 class InstanceLearner extends Learner { 29 private static final Comparator<Prediction> sComparator = new Comparator<Prediction>() { 30 public int compare(Prediction object1, Prediction object2) { 31 double score1 = object1.score; 32 double score2 = object2.score; 33 if (score1 > score2) { 34 return -1; 35 } else if (score1 < score2) { 36 return 1; 37 } else { 38 return 0; 39 } 40 } 41 }; 42 43 @Override classify(int sequenceType, int orientationType, float[] vector)44 ArrayList<Prediction> classify(int sequenceType, int orientationType, float[] vector) { 45 ArrayList<Prediction> predictions = new ArrayList<Prediction>(); 46 ArrayList<Instance> instances = getInstances(); 47 int count = instances.size(); 48 TreeMap<String, Double> label2score = new TreeMap<String, Double>(); 49 for (int i = 0; i < count; i++) { 50 Instance sample = instances.get(i); 51 if (sample.vector.length != vector.length) { 52 continue; 53 } 54 double distance; 55 if (sequenceType == GestureStore.SEQUENCE_SENSITIVE) { 56 distance = GestureUtils.minimumCosineDistance(sample.vector, vector, orientationType); 57 } else { 58 distance = GestureUtils.squaredEuclideanDistance(sample.vector, vector); 59 } 60 double weight; 61 if (distance == 0) { 62 weight = Double.MAX_VALUE; 63 } else { 64 weight = 1 / distance; 65 } 66 Double score = label2score.get(sample.label); 67 if (score == null || weight > score) { 68 label2score.put(sample.label, weight); 69 } 70 } 71 72 // double sum = 0; 73 for (String name : label2score.keySet()) { 74 double score = label2score.get(name); 75 // sum += score; 76 predictions.add(new Prediction(name, score)); 77 } 78 79 // normalize 80 // for (Prediction prediction : predictions) { 81 // prediction.score /= sum; 82 // } 83 84 Collections.sort(predictions, sComparator); 85 86 return predictions; 87 } 88 } 89