public PopulateVirtualWorld(String prefix) {
distribution = Configuration.getInt(prefix + "." + PAR_DISTRIB);
applicativeLayerId = Configuration.getPid(prefix + "." + APPLICATIVE_LAYER);
Globals.mapSize = Configuration.getLong(prefix + "." + MAP_SIZE);
Globals.zoneSize = Configuration.getLong(prefix + "." + ZONE_SIZE);
Globals.zoneNb = Configuration.getInt(prefix + "." + ZONE_NB);
Globals.outOfZoneNb = Configuration.getInt(prefix + "." + OUT_OF_ZONE_NB);
Globals.wSpeed = Configuration.getInt(prefix + "." + WSPEED);
Globals.tSpeed = Configuration.getInt(prefix + "." + TSPEED);
Globals.smallZoneSize = Configuration.getInt(prefix + "." + SMALL_ZONE_SIZE);
Globals.smallZoneNb = Configuration.getInt(prefix + "." + SMALL_ZONE_NB);
Globals.haltedToHalted = Configuration.getInt(prefix + "." + HTH);
Globals.haltedToTravelling = Configuration.getInt(prefix + "." + HTT);
Globals.haltedToWandering = Configuration.getInt(prefix + "." + HTW);
Globals.travellingToTravelling = Configuration.getInt(prefix + "." + TTT);
Globals.travellingToHalted = Configuration.getInt(prefix + "." + TTH);
Globals.travellingToWandering = Configuration.getInt(prefix + "." + TTW);
Globals.wanderingToWandering = Configuration.getInt(prefix + "." + WTW);
Globals.wanderingToTravelling = Configuration.getInt(prefix + "." + WTT);
Globals.wanderingToHalted = Configuration.getInt(prefix + "." + WTH);
Globals.changeWanderingDirection = Configuration.getInt(prefix + "." + CWD);
Globals.quiet = Configuration.getBoolean(prefix + "." + QUIET);
}
public static void main(String[] args) throws Exception {
String configName = args[0];
Configuration.setConfig(new ParsedProperties(configName));
// System.err.println("#Reading datasets from files");
File tFile = new File(Configuration.getString("trainingFile"));
File eFile = new File(Configuration.getString("evaluationFile"));
String dbReaderName = "gossipLearning.RecSysDataBaseReader";
seed = Configuration.getInt("SEED");
numOfClusters = Configuration.getInt("numOfClusters");
divErr = Configuration.getDouble("divErr");
numberOfRatings = Configuration.getDouble("numOfRatings");
numberOfAllUsers = Configuration.getDouble("numOfAllUsers");
classifierName = Configuration.getString("classifier", null);
prefix = "model";
modelName = Configuration.getString(prefix, null);
boolean isNormalize = Configuration.getBoolean("normalize");
int polyDegree = Configuration.getInt("polyDegree");
DataBaseReader dbReader = DataBaseReader.createDataBaseReader(dbReaderName, tFile, eFile);
// System.err.println("#Preprocessing datasets");
// pre-processing database
getNumberOfUsers = new double[dbReader.getTrainingSet().getNumberOfFeatures()];
getAverageRating = new double[dbReader.getTrainingSet().getNumberOfFeatures()];
getNumberOfLikeableUsers = new double[dbReader.getTrainingSet().getNumberOfFeatures()];
for (int i = 0; i < dbReader.getTrainingSet().size(); i++) {
for (VectorEntry e : dbReader.getTrainingSet().getInstance(i)) {
getNumberOfUsers[e.index]++;
getAverageRating[e.index] += e.value;
}
}
for (int i = 0; i < dbReader.getTrainingSet().getNumberOfFeatures(); i++) {
getAverageRating[i] =
getNumberOfUsers[i] == 0.0 ? 0.0 : getAverageRating[i] / getNumberOfUsers[i];
}
for (int i = 0; i < dbReader.getTrainingSet().size(); i++) {
for (VectorEntry e : dbReader.getTrainingSet().getInstance(i)) {
if (e.value > getAverageRating[e.index]) {
getNumberOfLikeableUsers[e.index]++;
}
}
}
// System.err.println("#Extracting features");
// feature extraction
InstanceHolder trainingSet = new InstanceHolder((int) numberOfRatings, 11);
for (int userID = 0; userID < dbReader.getTrainingSet().size(); userID++) {
for (VectorEntry e : dbReader.getTrainingSet().getInstance(userID)) {
int itemID = e.index;
double label = e.value;
double[] features = new double[11];
computeUserSpecificFeatureVectorPart(
dbReader.getTrainingSet().getInstance(userID), features);
computeItemSpecificFinalFeatureVector(itemID, features);
SparseVector instance = new SparseVector(features);
trainingSet.add(instance, label);
}
}
InstanceHolder testSet = new InstanceHolder((int) numberOfRatings, 11);
for (int userID = 0; userID < dbReader.getEvalSet().size(); userID++) {
for (VectorEntry e : dbReader.getEvalSet().getInstance(userID)) {
int itemID = e.index;
double label = e.value;
double[] features = new double[11];
computeUserSpecificFeatureVectorPart(
dbReader.getTrainingSet().getInstance(userID), features);
computeItemSpecificFinalFeatureVector(itemID, features);
SparseVector instance = new SparseVector(features);
testSet.add(instance, label);
}
}
File tout = new File("tr.dat");
File eout = new File("ev.dat");
trainingSet.writeToFile(tout);
testSet.writeToFile(eout);
DataBaseReader dbr =
DataBaseReader.createDataBaseReader("gossipLearning.DataBaseReader", tout, eout);
if (polyDegree > 1) {
// dbr.polynomize(polyDegree);
// dbr.writeToFile(new File("bookcrossing_fn_" + polyDegree + ".train"), new
// File("bookcrossing_fn_" + polyDegree + ".test"));
// System.exit(0);
}
if (isNormalize) {
dbr.standardize();
}
if (classifierName != null) {
weka(dbr.getTrainingSet(), dbr.getEvalSet());
// weka(trainingSet, testSet);
}
if (modelName != null) {
golf(dbr.getTrainingSet(), dbr.getEvalSet());
// golf(trainingSet, testSet);
}
tout.deleteOnExit();
eout.deleteOnExit();
}
