use of de.lmu.ifi.dbs.elki.distance.distancefunction.geo.LatLngDistanceFunction in project elki by elki-project.
the class GeoIndexing method main.
public static void main(String[] args) {
// Set the logging level to statistics:
LoggingConfiguration.setStatistics();
// Generate a random data set.
Random rand = new Random(0L);
// Note: ELKI has a nice data generator class, use that instead.
double[][] data = new double[100000][];
for (int i = 0; i < data.length; i++) {
data[i] = randomLatitudeLongitude(rand);
}
// Adapter to load data from an existing array.
DatabaseConnection dbc = new ArrayAdapterDatabaseConnection(data);
// Since the R-tree has so many options, it is a bit easier to configure it
// using the parameterization API, which handles defaults, instantiation,
// and additional constraint checks.
RStarTreeFactory<?> indexfactory = //
new ELKIBuilder<>(RStarTreeFactory.class).with(AbstractPageFileFactory.Parameterizer.PAGE_SIZE_ID, //
512).with(RStarTreeFactory.Parameterizer.BULK_SPLIT_ID, //
SortTileRecursiveBulkSplit.class).build();
// Create the database, and initialize it.
Database db = new StaticArrayDatabase(dbc, Arrays.asList(indexfactory));
// This will build the index of the database.
db.initialize();
// Relation containing the number vectors we put in above:
Relation<NumberVector> rel = db.getRelation(TypeUtil.NUMBER_VECTOR_FIELD);
// We can use this to identify rows of the input data below.
DBIDRange ids = (DBIDRange) rel.getDBIDs();
// For all indexes, dump their statistics.
for (It<Index> it = db.getHierarchy().iterDescendants(db).filter(Index.class); it.valid(); it.advance()) {
it.get().logStatistics();
}
// We use the WGS84 earth model, and "latitude, longitude" coordinates:
// This distance function returns meters.
LatLngDistanceFunction df = new LatLngDistanceFunction(WGS84SpheroidEarthModel.STATIC);
// k nearest neighbor query:
KNNQuery<NumberVector> knnq = QueryUtil.getKNNQuery(rel, df);
// Let's find the closest points to New York:
DoubleVector newYork = DoubleVector.wrap(new double[] { 40.730610, -73.935242 });
KNNList knns = knnq.getKNNForObject(newYork, 10);
// Iterate over all results.
System.out.println("Close to New York:");
for (DoubleDBIDListIter it = knns.iter(); it.valid(); it.advance()) {
// To kilometers
double km = it.doubleValue() / 1000;
System.out.println(rel.get(it) + " distance: " + km + " km row: " + ids.getOffset(it));
}
// Many other indexes will fail if we search close to the date line.
DoubleVector tuvalu = DoubleVector.wrap(new double[] { -7.4784205, 178.679924 });
knns = knnq.getKNNForObject(tuvalu, 10);
// Iterate over all results.
System.out.println("Close to Tuvalu:");
for (DoubleDBIDListIter it = knns.iter(); it.valid(); it.advance()) {
// To kilometers
double km = it.doubleValue() / 1000;
System.out.println(rel.get(it) + " distance: " + km + " km row: " + ids.getOffset(it));
}
// the distances to a few points in the data set.
for (It<Index> it = db.getHierarchy().iterDescendants(db).filter(Index.class); it.valid(); it.advance()) {
it.get().logStatistics();
}
}
Aggregations