@Override
public List<Classifier> buildClassifiers(User user, List<Sample> validSamples) {
Instances trainingSet =
new TrainingSetBuilder()
.setAttributes(user.getBssids())
.setClassAttribute(
"Location",
user.getLocations().stream().map(Location::getName).collect(Collectors.toList()))
.build("TrainingSet", validSamples.size());
// Create instances
validSamples.forEach(
sample -> {
Map<String, Integer> BSSIDLevelMap = getBSSIDLevelMap(sample);
Instance instance = new Instance(trainingSet.numAttributes());
for (Enumeration e = trainingSet.enumerateAttributes(); e.hasMoreElements(); ) {
Attribute attribute = (Attribute) e.nextElement();
String bssid = attribute.name();
int level = (BSSIDLevelMap.containsKey(bssid)) ? BSSIDLevelMap.get(bssid) : 0;
instance.setValue(attribute, level);
}
instance.setValue(trainingSet.classAttribute(), sample.getLocation());
instance.setDataset(trainingSet);
trainingSet.add(instance);
});
// Build classifiers
List<Classifier> classifiers = buildClassifiers(trainingSet);
return classifiers;
}
@Override
public String classify(User user, Sample sample) {
Instances trainingSet =
new TrainingSetBuilder()
.setAttributes(user.getBssids())
.setClassAttribute(
"Location",
user.getLocations().stream().map(Location::getName).collect(Collectors.toList()))
.build("TrainingSet", 1);
// Create instance
Map<String, Integer> BSSIDLevelMap = getBSSIDLevelMap(sample);
Instance instance = new Instance(trainingSet.numAttributes());
for (Enumeration e = trainingSet.enumerateAttributes(); e.hasMoreElements(); ) {
Attribute attribute = (Attribute) e.nextElement();
String bssid = attribute.name();
int level = (BSSIDLevelMap.containsKey(bssid)) ? BSSIDLevelMap.get(bssid) : 0;
instance.setValue(attribute, level);
}
if (sample.getLocation() != null)
instance.setValue(trainingSet.classAttribute(), sample.getLocation());
instance.setDataset(trainingSet);
trainingSet.add(instance);
int predictedClass = classify(fromBase64(user.getClassifiers()), instance);
return trainingSet.classAttribute().value(predictedClass);
}
private static void loadattributes(
String s,
Instance instclassificationpickup,
Instance instclassificationdropoff,
Instance instregressionpickup,
Instance instregressiondropoff) {
String[] attributes = s.split(",");
for (int i = 0; i < 8; i++) {
if (attributes[i].equals("?")) {
instclassificationpickup.setMissing(i);
instclassificationdropoff.setMissing(i);
instregressionpickup.setMissing(i);
instregressiondropoff.setMissing(i);
} else {
if ((i == 2) || (i == 3) || (i == 5)) {
instclassificationpickup.setValue(i, attributes[i]);
instclassificationdropoff.setValue(i, attributes[i]);
instregressionpickup.setValue(i, attributes[i]);
instregressiondropoff.setValue(i, attributes[i]);
} else {
instclassificationpickup.setValue(i, Double.parseDouble(attributes[i]));
instclassificationdropoff.setValue(i, Double.parseDouble(attributes[i]));
instregressionpickup.setValue(i, Double.parseDouble(attributes[i]));
instregressiondropoff.setValue(i, Double.parseDouble(attributes[i]));
}
}
}
instclassificationpickup.setMissing(8);
instclassificationdropoff.setMissing(8);
instregressionpickup.setMissing(8);
instregressiondropoff.setMissing(8);
}
private void addInstance(
final double min,
final double max,
final double mean,
final double stdDev,
final String label) {
Instance dataInstance = new Instance(NUMBER_OF_ATTRIBUTES);
dataInstance.setValue(minAttribute, min);
dataInstance.setValue(maxAttribute, max);
dataInstance.setValue(meanAttribute, mean);
dataInstance.setValue(stdDevAttribute, stdDev);
dataInstance.setValue(activityAttribute, label);
instances.add(dataInstance);
}
// 将一个文本数据转换成weka可以识别的Instance
private Instance makeInstance(String text, Instances data) {
Instance instance = new Instance(2);
Attribute messageAtt = data.attribute("Message");
instance.setValue(messageAtt, messageAtt.addStringValue(text));
instance.setDataset(data);
return instance;
}
private Instance instanceFromStream(StreamElement data) {
try {
Instance i = new Instance(data.getFieldNames().length);
for (int j = 0; j < data.getFieldNames().length; j++) {
i.setValue(j, ((Double) data.getData()[j]));
}
// scaling specific to opensense data!! should be put in the parameters?
i.setValue(0, i.value(0) / 1400.0);
i.setValue(2, i.value(2) / 50);
i.setValue(3, i.value(3) / 100.0);
i.setValue(4, i.value(4) / 100.0 - 4);
return i;
} catch (Exception e) {
return null;
}
}
/**
* Input an instance for filtering.
*
* @param instance the input instance
* @return true if the filtered instance may now be collected with output().
* @throws Exception if the input format was not set or the date format cannot be parsed
*/
public boolean input(Instance instance) throws Exception {
if (getInputFormat() == null) {
throw new IllegalStateException("No input instance format defined");
}
if (m_NewBatch) {
resetQueue();
m_NewBatch = false;
}
Instance newInstance = (Instance) instance.copy();
int index = m_AttIndex.getIndex();
if (!newInstance.isMissing(index)) {
double value = instance.value(index);
try {
// Format and parse under the new format to force any required
// loss in precision.
value = m_OutputAttribute.parseDate(m_OutputAttribute.formatDate(value));
} catch (ParseException pe) {
throw new RuntimeException("Output date format couldn't parse its own output!!");
}
newInstance.setValue(index, value);
}
push(newInstance);
return true;
}
private void createLimbInstance(
Skeleton skeleton, int limb, int[] limb_about, List<Attribute> limb_att) {
int att = 1;
Instance ist = new DenseInstance((SPATIAL_COORD * limb_about.length) + 1);
limb_att = classifier.getLimbFeatureDataSet(limb);
for (int i = 0; i < limb_about.length; i++) {
if (limb_about[i] != TORSO) {
/*
* il punto di riferimento per tutti i joint (meno che per il
* TORSO), ossia il centro degli assi che descrivono lo spazio
* in cui si possono muovere è il TORSO stesso.
*/
Point3D<Float> joint = skeleton.getJoint(SKELETON_JOINT(limb_about[i])).getPosition();
ist.setValue((Attribute) limb_att.get(att), (joint.getX() - x_torso));
ist.setValue((Attribute) limb_att.get(att + 1), (joint.getY() - y_torso));
ist.setValue((Attribute) limb_att.get(att + 2), (joint.getZ() - z_torso));
att += 3;
} else {
/*
* il punto di riferimento per il BODY invece è il punto con
* coordinate (0, -Ytorso, 0), e cioè, il punto sul pavimento
* esattamente perpendicolare al TORSO stesso
*/
Point3D<Float> left_foot = skeleton.getJoint(JointType.LEFT_FOOT).getPosition();
Point3D<Float> right_foot = skeleton.getJoint(JointType.RIGHT_FOOT).getPosition();
// save the two probably torso's new coordinates
float torso_footSX = y_torso - left_foot.getY();
float torso_footDX = y_torso - right_foot.getY();
float y_torso_rifFoot = (torso_footSX > torso_footDX) ? torso_footSX : torso_footDX;
ist.setValue((Attribute) limb_att.get(att), 0);
ist.setValue((Attribute) limb_att.get(att + 1), y_torso_rifFoot);
ist.setValue((Attribute) limb_att.get(att + 2), 0);
att += 3;
}
}
skeletonIstances.add(ist);
}
protected void imputeMedian(Instances instances) {
Attribute indicator = instances.attribute(ATTNAME_INDICATOR);
for (int i = 0; i < instances.numInstances(); ++i) {
Instance current = instances.get(i);
current.setValue(indicator, 0.0); // 0.0 means "false"
for (int j = 0; j < instances.numAttributes(); ++j) {
if (instances.attribute(j).isNumeric() == false) {
continue;
}
if (Utils.isMissingValue(current.value(j))) {
current.setValue(j, medians[j]);
current.setValue(indicator, 1.0);
imputations[j] += 1;
}
}
}
}
/**
* Input an instance for filtering. The instance is processed and made available for output
* immediately.
*
* @param instance the input instance
* @return true if the filtered instance may now be collected with output().
* @throws IllegalStateException if no input format has been set.
*/
public boolean input(Instance instance) {
if (getInputFormat() == null) {
throw new IllegalStateException("No input instance format defined");
}
if (m_NewBatch) {
resetQueue();
m_NewBatch = false;
}
Instance newInstance = (Instance) instance.copy();
if ((int) newInstance.value(m_AttIndex.getIndex()) == m_SecondIndex.getIndex()) {
newInstance.setValue(m_AttIndex.getIndex(), (double) m_FirstIndex.getIndex());
} else if ((int) newInstance.value(m_AttIndex.getIndex()) > m_SecondIndex.getIndex()) {
newInstance.setValue(m_AttIndex.getIndex(), newInstance.value(m_AttIndex.getIndex()) - 1);
}
push(newInstance);
return true;
}
// create an instance in the format of CreateAppInsertIntoVm.arff
// eval can either be a Instance - MissingValue or the evaluation value
private Instance createInstance(double eval, VirtualMachine vm) {
Instance instance = new Instance(64);
LinkedList<Integer> cpuallhist = vm.getCpuAllocationHistory(10);
LinkedList<Integer> cpuusehist = vm.getCpuUsageHistory(10);
LinkedList<Integer> memallhist = vm.getMemoryAllocationHistory(10);
LinkedList<Integer> memusehist = vm.getMemoryUsageHistory(10);
LinkedList<Integer> storageallhist = vm.getStorageAllocationHistory(10);
LinkedList<Integer> storageusehist = vm.getStorageUsageHistory(10);
// CPU/Memory/Storage - Allocation history before the new vm was created
for (int i = 0; i < 10; i++) {
// cpu allocation
instance.setValue(getKnowledgeBase().attribute(i), clusterValue(cpuallhist.get(i)));
// cpu usage
instance.setValue(getKnowledgeBase().attribute(i + 10), clusterValue(cpuusehist.get(i)));
// memory allocation
instance.setValue(getKnowledgeBase().attribute(i + 20), clusterValue(memallhist.get(i)));
// memory usage
instance.setValue(getKnowledgeBase().attribute(i + 30), clusterValue(memusehist.get(i)));
// storage allocation
instance.setValue(getKnowledgeBase().attribute(i + 40), clusterValue(storageallhist.get(i)));
// storage usage
instance.setValue(getKnowledgeBase().attribute(i + 50), clusterValue(storageusehist.get(i)));
}
// SLAs
// CPU
instance.setValue(getKnowledgeBase().attribute(60), clusterValue(app.getCpu()));
// Memory
instance.setValue(getKnowledgeBase().attribute(61), clusterValue(app.getMemory()));
// Storage
instance.setValue(getKnowledgeBase().attribute(62), clusterValue(app.getStorage()));
// Evaluation
instance.setValue(getKnowledgeBase().attribute(63), eval);
return instance;
}
/**
* method to set a new value
*
* @param r random function
* @param numOfValues
* @param instance
* @param useMissing
*/
private void changeValueRandomly(
Random r, int numOfValues, int indexOfAtt, Instance instance, boolean useMissing) {
int currValue;
// get current value
// if value is missing set current value to number of values
// whiche is the highest possible value plus one
if (instance.isMissing(indexOfAtt)) {
currValue = numOfValues;
} else {
currValue = (int) instance.value(indexOfAtt);
}
// with only two possible values it is easier
if ((numOfValues == 2) && (!instance.isMissing(indexOfAtt))) {
instance.setValue(indexOfAtt, (double) ((currValue + 1) % 2));
} else {
// get randomly a new value not equal to the current value
// if missing values are used as values they must be treated
// in a special way
while (true) {
int newValue;
if (useMissing) {
newValue = (int) (r.nextDouble() * (double) (numOfValues + 1));
} else {
newValue = (int) (r.nextDouble() * (double) numOfValues);
}
// have we found a new value?
if (newValue != currValue) {
// the value 1 above the highest possible value (=numOfValues)
// is used as missing value
if (newValue == numOfValues) {
instance.setMissing(indexOfAtt);
} else {
instance.setValue(indexOfAtt, (double) newValue);
}
break;
}
}
}
}
/**
* Input an instance for filtering. The instance is processed and made available for output
* immediately.
*
* @param instance the input instance
* @return true if the filtered instance may now be collected with output().
* @throws IllegalStateException if no input structure has been defined.
*/
@Override
public boolean input(Instance instance) {
if (getInputFormat() == null) {
throw new IllegalStateException("No input instance format defined");
}
if (m_NewBatch) {
resetQueue();
m_NewBatch = false;
}
Instance newInstance = (Instance) instance.copy();
if (!newInstance.isMissing(m_AttIndex.getIndex())) {
if ((int) newInstance.value(m_AttIndex.getIndex()) == m_SecondIndex.getIndex()) {
newInstance.setValue(m_AttIndex.getIndex(), m_FirstIndex.getIndex());
} else if ((int) newInstance.value(m_AttIndex.getIndex()) == m_FirstIndex.getIndex()) {
newInstance.setValue(m_AttIndex.getIndex(), m_SecondIndex.getIndex());
}
}
push(newInstance, false); // No need to copy
return true;
}
/**
* classifies a tweet
*
* @param stringa the tweet to classify
* @return the tweet polarity
*/
public double classifyDouble(String stringa)
throws FileNotFoundException, IOException, Exception {
String string_new;
Preprocesser pre = new Preprocesser();
string_new = pre.preprocessDocument(stringa);
String tmp = "";
StringTokenizer st = new StringTokenizer(string_new, " ");
// Instances unlabeled = new Instances(new BufferedReader(new
// FileReader("D:/prova.arff")));
Instances unlabeled = new Instances(_train, 1);
Instance inst = new Instance(unlabeled.numAttributes());
// load unlabeled data
inst.setDataset(unlabeled);
int j = 0;
while (j < unlabeled.numAttributes()) {
inst.setValue(j, "0");
j++;
}
while (st.hasMoreTokens()) {
tmp = st.nextToken();
if (unlabeled.attribute(tmp) != null) inst.setValue(unlabeled.attribute(tmp), "1");
}
unlabeled.add(inst);
// set class attribute
unlabeled.setClassIndex(unlabeled.numAttributes() - 1);
// create copy
Instances labeled = new Instances(unlabeled);
// label instances
for (int i = 0; i < unlabeled.numInstances(); i++) {
double clsLabel = _cl.classifyInstance(unlabeled.instance(i));
labeled.instance(i).setClassValue(clsLabel);
}
// return labeled.instance(0).stringValue(unlabeled.numAttributes()-1);
// System.out.println("weight: " + labeled.instance(0).weight());
return labeled.instance(0).classValue();
}
@Override
public Instance nextInstance() {
InstancesHeader header = getHeader();
Instance inst = new DenseInstance(header.numAttributes());
inst.setDataset(header);
int waveform = this.instanceRandom.nextInt(NUM_CLASSES);
int choiceA = 0, choiceB = 0;
switch (waveform) {
case 0:
choiceA = 0;
choiceB = 1;
break;
case 1:
choiceA = 0;
choiceB = 2;
break;
case 2:
choiceA = 1;
choiceB = 2;
break;
}
double multiplierA = this.instanceRandom.nextDouble();
double multiplierB = 1.0 - multiplierA;
for (int i = 0; i < NUM_BASE_ATTRIBUTES; i++) {
inst.setValue(
this.numberAttribute[i],
(multiplierA * hFunctions[choiceA][i])
+ (multiplierB * hFunctions[choiceB][i])
+ this.instanceRandom.nextGaussian());
}
if (this.addNoiseOption.isSet()) {
for (int i = NUM_BASE_ATTRIBUTES; i < TOTAL_ATTRIBUTES_INCLUDING_NOISE; i++) {
inst.setValue(this.numberAttribute[i], this.instanceRandom.nextGaussian());
}
}
inst.setClassValue(waveform);
return inst;
}
private static Instance createInstance(double[] ds, String string, Instances instances) {
int l = ds.length;
// Create the instance
Instance instance = new Instance(1 + l);
instance.setDataset(instances);
for (int i = 1; i <= l; i++) {
instance.setValue(i, ds[i - 1]);
}
if (string == null) {
instance.setClassMissing();
} else {
instance.setClassValue(string);
}
return instance;
}
/**
* Call this function to predict the class of an instance once a classification model has been
* built with the buildClassifier call.
*
* @param i The instance to classify.
* @return A double array filled with the probabilities of each class type.
* @throws Exception if can't classify instance.
*/
public double[] distributionForInstance(Instance i) throws Exception {
// Make a copy of the instance so that it isn't modified
m_currentInstance = (Instance) i.copy();
if (m_normalizeAttributes) {
for (int noa = 0; noa < m_currentInstance.dataset().numAttributes(); noa++) {
if (noa != m_currentInstance.classIndex()) {
if (m_attributeRanges[noa] != 0) {
m_currentInstance.setValue(
noa,
(m_currentInstance.value(noa) - m_attributeBases[noa]) / m_attributeRanges[noa]);
} else {
m_currentInstance.setValue(noa, m_currentInstance.value(noa) - m_attributeBases[noa]);
}
}
}
}
resetNetwork();
// since all the output values are needed.
// They are calculated manually here and the values collected.
double[] theArray = new double[m_numClasses];
double count = 0;
for (int noa = 0; noa < m_numClasses; noa++) {
theArray[noa] = m_outputs[noa].outputValue(true);
count += theArray[noa];
}
if (count <= 0) {
return m_ZeroR.m_Counts;
}
return theArray;
}
/**
* Transform - turn [y1,y2,y3,x1,x2] into [y1,y2,x1,x2].
*
* @return transformed Instance
*/
public Instance transform(Instance x, double ypred[]) throws Exception {
x = (Instance) x.copy();
int L = x.classIndex();
int L_c = (paY.length + 1);
x.setDataset(null);
for (int j = 0; j < (L - L_c); j++) {
x.deleteAttributeAt(0);
}
for (int pa : paY) {
// System.out.println("x_["+map[pa]+"] <- "+ypred[pa]);
x.setValue(map[pa], ypred[pa]);
}
x.setDataset(T);
x.setClassMissing();
return x;
}
@Override
public boolean evaluate() {
if (curInstance == null) {
curInstance = createInstance(0, selectedVm); // create a Instance with the past values
}
if (app.getSuspendedTicks() > 0
|| app.getVm().getSuspendedTicks() > 0
|| app.getVm().getPm().getSuspendedTicks() > 0) {
return false;
} else if (waitForEvaluation > 0) {
waitForEvaluation--;
return false;
} else {
// System.out.println("APP - Running Ticks: " + app.getRunningTicks());
LinkedList<Integer> cpuusagehist = selectedVm.getPm().getCpuUsageHistory(10);
LinkedList<Integer> memusagehist = selectedVm.getPm().getMemoryUsageHistory(10);
LinkedList<Integer> storageusagehist = selectedVm.getPm().getStorageUsageHistory(10);
double evaluation =
(255
- calculateUsageRatio(cpuusagehist, 85)
- calculateUsageRatio(memusagehist, 85)
- calculateUsageRatio(storageusagehist, 85))
/ 255;
// subtract SLA Violations
evaluation -=
(app.getCpuSlaErrorcount() + app.getMemorySlaErrorcount() + app.getStorageSlaErrorcount())
/ 10;
// minimum of 0
// evaluation = Math.max(0, evaluation);
Monitor.getInstance().logExecution(app, this, evaluation, this.globalTickExecution);
curInstance.setValue(getKnowledgeBase().attribute(63), evaluation);
this.setLocalEvaluation(evaluation);
getKnowledgeBase().add(curInstance);
}
return true;
}
@Override
public void setAttributeValue(Attribute attribute, Instance instance, Patient patient) {
RegaService rs = ruleService.getRegaService();
DataService ds = ruleService.getDataService();
List<TestResult> postTreatmentCD4Counts = rs.getPostTreatmentCD4Counts(patient);
Double leastSquaresRegressionGradient = ds.getLeastSquaresFitGradient(postTreatmentCD4Counts);
logger.info(
"PTCD4 #: "
+ postTreatmentCD4Counts.size()
+ ", Gradient: "
+ leastSquaresRegressionGradient);
if (leastSquaresRegressionGradient != null)
instance.setValue(attribute, leastSquaresRegressionGradient);
}
private Instance getFVSFilteredInstance(
Instances output, Instance old_inst, List<List<Value>> list, Double[] substitution) {
double[] oldValues = old_inst.toDoubleArray();
Instance instance = new Instance(old_inst);
// Change with value that is available
for (int i = 0; i < oldValues.length - 1; i++) {
// System.out.println(oldValues[i]);
// System.out.println(list.get(i));
// System.out.println("############################");
// If list doesn't contain, then delete
Value v = new Value(oldValues[i]);
int idx = list.get(i).indexOf(v);
// If not found in the index
if (idx == -1) {
// Change with substitution
instance.setValue(i, substitution[i]);
// Change into missing
// instance.setMissing(i);
}
}
return instance;
}
private static Instance makeInstance(Instances instances, String inputLine) {
inputLine = inputLine.trim();
// We need to store the lastName as well...
String[] parts = inputLine.split("\\s+");
String label = parts[0];
String firstName = parts[1].toLowerCase();
String lastName = parts[2].toLowerCase();
Instance instance = new Instance(features.length + 1);
instance.setDataset(instances);
Set<String> feats = new HashSet<String>();
/*
feats.add("firstName0=" + firstName.charAt(0));
feats.add("firstNameN=" + firstName.charAt(firstName.length() - 1));
*/
for (int f = 0; f < 9; f++) {
if (firstName.length() > f) feats.add("firstName" + f + "=" + firstName.charAt(f));
}
for (int l = 0; l < 9; l++) {
if (lastName.length() > l) feats.add("lastName" + l + "=" + lastName.charAt(l));
}
/////////////////////////////////////////////////////////////////
for (int featureId = 0; featureId < features.length; featureId++) {
Attribute att = instances.attribute(features[featureId]);
String name = att.name();
String featureLabel;
if (feats.contains(name)) {
featureLabel = "1";
} else featureLabel = "0";
instance.setValue(att, featureLabel);
}
instance.setClassValue(label);
return instance;
}
/**
* Create an Instance of the same type as the Instances object you are searching in.
*
* @param p - a 3D point
* @param dataset - the dataset you are searching in, which was used to build the KDTree
* @return an Instance that the nearest neighbor can be found for
*/
public Instance createInstance(Integer points[], final Instances dataset) {
// Create numeric attributes "x" and "y" and "z"
Attribute attr[] = new Attribute[10];
for (int i = 0; i < points.length; i++) {
attr[i] = new Attribute("p" + i);
attr[i] = dataset.attribute(i);
}
// Create vector of the above attributes
FastVector attributes = new FastVector(points.length);
for (int i = 0; i < points.length; i++) attributes.addElement(attr[i]);
// Create empty instance with X attribute values
Instance inst = new Instance(points.length);
// Set instance's values for the attributes "x", "y", and "z"
for (int i = 0; i < points.length; i++) inst.setValue(attr[i], points[i]);
// Set instance's dataset to be the dataset "points1"
inst.setDataset(dataset);
return inst;
}
public void buildClusterer(ArrayList<String> seqDB, double[][] sm) {
seqList = seqDB;
this.setSimMatrix(sm);
Attribute seqString = new Attribute("sequence", (FastVector) null);
FastVector attrInfo = new FastVector();
attrInfo.addElement(seqString);
Instances data = new Instances("data", attrInfo, 0);
for (int i = 0; i < seqList.size(); i++) {
Instance currentInst = new Instance(1);
currentInst.setDataset(data);
currentInst.setValue(0, seqList.get(i));
data.add(currentInst);
}
try {
buildClusterer(data);
} catch (Exception e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
@Override
public void setAttributeValue(Attribute attribute, Instance instance, Patient patient) {
RegaService rs = ruleService.getRegaService();
PatientAttributeValue pav =
rs.getPatientAttributeValue(
patient, ClinicalAttributeFactory.CLINICAL_REGA_ATTRIBUTE_RECENT_BLOOD_HB);
if (pav == null) return; // The is no recent blood hb for this patient
String strVal = pav.getValue();
Double dVal;
if (strVal.equals("N/A")) {
return; // There is no value
} else if (strVal.contains("/")) {
dVal = Double.parseDouble(strVal.replace('/', '.'));
} else {
dVal = Double.parseDouble(strVal);
}
logger.info("Recent Blood HB for " + patient.getPatientId() + " is " + dVal);
instance.setValue(attribute, dVal);
}
/**
* Input an instance for filtering. The instance is processed and made available for output
* immediately.
*
* @param instance the input instance.
* @return true if the filtered instance may now be collected with output().
* @throws IllegalStateException if no input structure has been defined.
*/
public boolean input(Instance instance) {
if (getInputFormat() == null) {
throw new IllegalStateException("No input instance format defined");
}
if (m_NewBatch) {
resetQueue();
m_NewBatch = false;
}
if (isOutputFormatDefined()) {
Instance newInstance = (Instance) instance.copy();
// make sure that we get the right indexes set for the converted
// string attributes when operating on a second batch of instances
for (int i = 0; i < newInstance.numAttributes(); i++) {
if (newInstance.attribute(i).isString()
&& !newInstance.isMissing(i)
&& m_AttIndices.isInRange(i)) {
Attribute outAtt = getOutputFormat().attribute(newInstance.attribute(i).name());
String inVal = newInstance.stringValue(i);
int outIndex = outAtt.indexOfValue(inVal);
if (outIndex < 0) {
newInstance.setMissing(i);
} else {
newInstance.setValue(i, outIndex);
}
}
}
push(newInstance);
return true;
}
bufferInput(instance);
return false;
}
/**
* Processes the given data (may change the provided dataset) and returns the modified version.
* This method is called in batchFinished(). This implementation only calls process(Instance) for
* each instance in the given dataset.
*
* @param instances the data to process
* @return the modified data
* @throws Exception in case the processing goes wrong
* @see #batchFinished()
*/
protected Instances process(Instances instances) throws Exception {
Instances result;
Instance instOld;
Instance instNew;
int i;
int n;
double[] values;
int numAttNew;
int numAttOld;
if (!isFirstBatchDone()) computeThresholds(instances);
result = getOutputFormat();
numAttOld = instances.numAttributes();
numAttNew = result.numAttributes();
for (n = 0; n < instances.numInstances(); n++) {
instOld = instances.instance(n);
values = new double[numAttNew];
System.arraycopy(instOld.toDoubleArray(), 0, values, 0, numAttOld);
// generate new instance
instNew = new Instance(1.0, values);
instNew.setDataset(result);
// per attribute?
if (!getDetectionPerAttribute()) {
// outlier?
if (isOutlier(instOld)) instNew.setValue(m_OutlierAttributePosition[0], 1);
// extreme value?
if (isExtremeValue(instOld)) {
instNew.setValue(m_OutlierAttributePosition[0] + 1, 1);
// tag extreme values also as outliers?
if (getExtremeValuesAsOutliers()) instNew.setValue(m_OutlierAttributePosition[0], 1);
}
} else {
for (i = 0; i < m_AttributeIndices.length; i++) {
// non-numeric attribute?
if (m_AttributeIndices[i] == NON_NUMERIC) continue;
// outlier?
if (isOutlier(instOld, m_AttributeIndices[i]))
instNew.setValue(m_OutlierAttributePosition[i], 1);
// extreme value?
if (isExtremeValue(instOld, m_AttributeIndices[i])) {
instNew.setValue(m_OutlierAttributePosition[i] + 1, 1);
// tag extreme values also as outliers?
if (getExtremeValuesAsOutliers()) instNew.setValue(m_OutlierAttributePosition[i], 1);
}
// add multiplier?
if (getOutputOffsetMultiplier())
instNew.setValue(
m_OutlierAttributePosition[i] + 2,
calculateMultiplier(instOld, m_AttributeIndices[i]));
}
}
// copy possible strings, relational values...
copyValues(instNew, false, instOld.dataset(), getOutputFormat());
// add to output
result.add(instNew);
}
return result;
}
public static void main(String[] args) throws Exception {
// NaiveBayesSimple nb = new NaiveBayesSimple();
// BufferedReader br_train = new BufferedReader(new FileReader("src/train.arff.txt"));
// String s = null;
// long st_time = System.currentTimeMillis();
// Instances inst_train = new Instances(br_train);
// System.out.println(inst_train.numAttributes());
// inst_train.setClassIndex(inst_train.numAttributes()-1);
// System.out.println("train time"+(System.currentTimeMillis()-st_time));
// NaiveBayes nb1 = new NaiveBayes();
// nb1.buildClassifier(inst_train);
// br_train.close();
long st_time = System.currentTimeMillis();
st_time = System.currentTimeMillis();
Classifier classifier = (Classifier) SerializationHelper.read("NaiveBayes.model");
// BufferedReader br_test = new BufferedReader(new FileReader("src/test.arff.txt"));
// Instances inst_test = new Instances(br_test);
// inst_test.setClassIndex(inst_test.numAttributes()-1);
// System.out.println("test time"+(System.currentTimeMillis()-st_time));
//
ArffLoader testLoader = new ArffLoader();
testLoader.setSource(new File("src/test.arff"));
testLoader.setRetrieval(Loader.BATCH);
Instances testDataSet = testLoader.getDataSet();
Attribute testAttribute = testDataSet.attribute("class");
testDataSet.setClass(testAttribute);
int correct = 0;
int incorrect = 0;
FastVector attInfo = new FastVector();
attInfo.addElement(new Attribute("Id"));
attInfo.addElement(new Attribute("Category"));
Instances outputInstances = new Instances("predict", attInfo, testDataSet.numInstances());
Enumeration testInstances = testDataSet.enumerateInstances();
int index = 1;
while (testInstances.hasMoreElements()) {
Instance instance = (Instance) testInstances.nextElement();
double classification = classifier.classifyInstance(instance);
Instance predictInstance = new Instance(outputInstances.numAttributes());
predictInstance.setValue(0, index++);
predictInstance.setValue(1, (int) classification + 1);
outputInstances.add(predictInstance);
}
System.out.println("Correct Instance: " + correct);
System.out.println("IncCorrect Instance: " + incorrect);
double accuracy = (double) (correct) / (double) (correct + incorrect);
System.out.println("Accuracy: " + accuracy);
CSVSaver predictedCsvSaver = new CSVSaver();
predictedCsvSaver.setFile(new File("predict.csv"));
predictedCsvSaver.setInstances(outputInstances);
predictedCsvSaver.writeBatch();
System.out.println("Prediciton saved to predict.csv");
}
public static void classify(Instances train, File file) throws Exception {
FastVector atts = new FastVector();
String[] classes = {"classical", "hiphop", "pop", "rock"};
double[] val;
FastVector attValsRel = new FastVector();
// Setting attributes for the test data
Attribute attributeZero = new Attribute("Zero_Crossings");
atts.addElement(attributeZero);
Attribute attributeLPC = new Attribute("LPC");
atts.addElement(attributeLPC);
Attribute attributeCentroid = new Attribute("Spectral_Centroid");
atts.addElement(attributeCentroid);
Attribute attributeRollOff = new Attribute("Spectral_Rolloff_Point");
atts.addElement(attributeRollOff);
Attribute attributePeakDetection = new Attribute("Peak_Detection");
atts.addElement(attributePeakDetection);
Attribute attributeStrongestBeat = new Attribute("Strongest_Beat");
atts.addElement(attributeStrongestBeat);
Attribute attributeBeatSum = new Attribute("Beat_Sum");
atts.addElement(attributeBeatSum);
Attribute attributeRMS = new Attribute("RMS");
atts.addElement(attributeRMS);
Attribute attributeConstantQ = new Attribute("ConstantQ");
atts.addElement(attributeConstantQ);
Attribute attributeMFT = new Attribute("MagnitudeFFT");
atts.addElement(attributeMFT);
Attribute attributeMFCC = new Attribute("MFCC");
atts.addElement(attributeMFCC);
for (int i = 0; i < classes.length; i++) attValsRel.addElement(classes[i]);
atts.addElement(new Attribute("class", attValsRel));
// Adding instances to the test dataset
Instances test = new Instances("AudioSamples", atts, 0);
val = makeData(file, null, attValsRel, test.numAttributes());
Instance instance = new Instance(test.numAttributes());
instance.setValue(attributeZero, val[0]);
instance.setValue(attributeLPC, val[1]);
instance.setValue(attributeCentroid, val[2]);
instance.setValue(attributeRollOff, val[3]);
instance.setValue(attributePeakDetection, val[4]);
instance.setValue(attributeStrongestBeat, val[5]);
instance.setValue(attributeBeatSum, val[6]);
instance.setValue(attributeRMS, val[7]);
instance.setValue(attributeConstantQ, val[8]);
instance.setValue(attributeMFT, val[9]);
instance.setValue(attributeMFCC, val[10]);
test.add(instance);
// Setting the class attribute
test.setClassIndex(test.numAttributes() - 1);
System.out.println(test);
// Trainging the classifier with train dataset
classifier = new FilteredClassifier();
classifier.buildClassifier(train);
// Classifying the test data with the train data
for (int i = 0; i < test.numInstances(); i++) {
double pred = classifier.classifyInstance(test.instance(i));
System.out.println("===== Classified instance =====");
System.out.println("Class predicted: " + test.classAttribute().value((int) pred));
}
}
public void updateTransform(Instance t_, double ypred[]) throws Exception {
for (int pa : this.paY) {
t_.setValue(this.map[pa], ypred[pa]);
}
}