List of usage examples for org.apache.commons.math.stat.descriptive.rank Percentile evaluate
public double evaluate(final double[] values, final double p)
pth percentile of the values in the values array. From source file:com.discursive.jccook.math.StatExample.java
public static void main(String[] args) { double[] values = new double[] { 2.3, 5.4, 6.2, 7.3, 23.3 }; System.out.println("min: " + StatUtils.min(values)); System.out.println("max: " + StatUtils.max(values)); System.out.println("mean: " + StatUtils.mean(values)); System.out.println("product: " + StatUtils.product(values)); System.out.println("sum: " + StatUtils.sum(values)); System.out.println("variance: " + StatUtils.variance(values)); // Measures from previous example Min min = new Min(); System.out.println("min: " + min.evaluate(values)); Max max = new Max(); System.out.println("max: " + max.evaluate(values)); Mean mean = new Mean(); System.out.println("mean: " + mean.evaluate(values)); Product product = new Product(); System.out.println("product: " + product.evaluate(values)); Sum sum = new Sum(); System.out.println("sum: " + sum.evaluate(values)); Variance variance = new Variance(); System.out.println("variance: " + variance.evaluate(values)); // New measures Percentile percentile = new Percentile(); System.out.println("80 percentile value: " + percentile.evaluate(values, 80.0)); GeometricMean geoMean = new GeometricMean(); System.out.println("geometric mean: " + geoMean.evaluate(values)); StandardDeviation stdDev = new StandardDeviation(); System.out.println("standard dev: " + stdDev.evaluate(values)); Skewness skewness = new Skewness(); System.out.println("skewness: " + skewness.evaluate(values)); Kurtosis kurtosis = new Kurtosis(); System.out.println("kurtosis: " + kurtosis.evaluate(values)); }
From source file:com.eviware.loadui.util.statistics.ValueStatistics.java
public synchronized Map<String, Number> getData(long timestamp) { long max = 0; long min = Long.MAX_VALUE; long sum = 0; for (Iterator<DataPoint> it = dataPoints.iterator(); it.hasNext();) { DataPoint dataPoint = it.next(); if (dataPoint.timestamp < timestamp - period && period > 0) it.remove();//from www. java 2 s . c om else { sum += dataPoint.value; max = Math.max(max, dataPoint.value); min = Math.min(min, dataPoint.value); } } int count = dataPoints.size(); double avg = count > 0 ? (double) sum / count : 0; double stdDev = 0; double[] dataSet = new double[count]; if (count > 0) { int i = 0; for (DataPoint dataPoint : dataPoints) { dataSet[i] = dataPoint.value; i++; stdDev += Math.pow(dataPoint.value - avg, 2); } stdDev = Math.sqrt(stdDev / count); } double tps = 0; long vps = 0; long duration = 0; if (count >= 2) { int samples = 0; long earliest = timestamp - snapshotLength; DataPoint point = null; while (++samples < count) { point = dataPoints.get(count - samples); vps += point.value; if (point.timestamp < earliest) break; } long timeDelta = timestamp - Preconditions.checkNotNull(point).timestamp; timeDelta = timeDelta == 0 ? 1000 : timeDelta; vps = vps * 1000 / timeDelta; tps = (samples - 1) * 1000.0 / timeDelta; duration = dataPoints.get(count - 1).timestamp - dataPoints.get(0).timestamp; } Percentile perc = new Percentile(90); double percentile = perc.evaluate(dataSet, 90); return new ImmutableMap.Builder<String, Number>() // .put("Max", max) // .put("Min", min == Long.MAX_VALUE ? 0L : min) // .put("Avg", avg) // .put("Sum", sum) // .put("Std-Dev", stdDev) // .put("Tps", tps) // .put("Avg-Tps", duration > 0L ? 1000L * count / duration : 0) // .put("Vps", vps) // .put("Avg-Vps", duration > 0L ? 1000L * sum / duration : 0) // .put("Percentile", percentile) // .put("AvgResponseSize", 1000L * sum / (dataPoints.size() == 0 ? 1 : dataPoints.size())) // .build(); }
From source file:fr.ens.transcriptome.corsen.util.StatTest.java
public void testFirstQuartile() { Percentile percentile = new Percentile(); for (int i = 0; i < 1000; i++) { List<DataDouble> list = generate(); assertEquals(percentile.evaluate(Stats.toDouble(list), 25.0), Stats.firstQuartile(list)); }//from www . j a v a2 s. com }
From source file:fr.ens.transcriptome.corsen.util.StatTest.java
public void testThirdQuartile() { Percentile percentile = new Percentile(); for (int i = 0; i < 1000; i++) { List<DataDouble> list = generate(); assertEquals(percentile.evaluate(Stats.toDouble(list), 75.0), Stats.thirdQuartile(list)); }/* w w w . j a va2 s .c o m*/ }
From source file:fr.ens.transcriptome.corsen.util.StatTest.java
public void testPercentile() { Percentile percentile = new Percentile(); for (int j = 0; j < 10; j++) { for (int i = 0; i < 98; i++) { List<DataDouble> list = generate(); double p = random.nextInt(99) + 1; assertEquals(percentile.evaluate(Stats.toDouble(list), p), Stats.percentile(list, p)); }//from ww w . j a v a 2 s . com } }
From source file:edu.cornell.med.icb.goby.modes.CompactFileStatsMode.java
/** * Print statistics about a reads file in the Goby compact form. * * @param file The file to display statistics about * @throws IOException if the file cannot be read *//* w w w. ja v a 2 s . c om*/ private void describeCompactReads(final File file) throws IOException { stream.printf("Compact reads filename = %s%n", file); // keep the read lengths for computing quantiles final DoubleArrayList readLengths = new DoubleArrayList(); int minLength = Integer.MAX_VALUE; int maxLength = Integer.MIN_VALUE; int numberOfIdentifiers = 0; int numberOfDescriptions = 0; int numberOfSequences = 0; int numberOfSequencePairs = 0; int numberOfQualityScores = 0; int numberOfQualityScorePairs = 0; long totalReadLength = 0; long totalReadLengthPair = 0; final DistinctIntValueCounterBitSet allQueryIndices = new DistinctIntValueCounterBitSet(); ReadsReader reader = null; boolean checkedForPaired = false; try { final long size = file.length(); reader = new ReadsReader(FileUtils.openInputStream(file)); for (final Reads.ReadEntry entry : reader) { final int readLength = entry.getReadLength(); for (int i = 0; i < entry.getMetaDataCount(); i++) { Reads.MetaData metaData = entry.getMetaData(i); stream.printf("meta-data key=%s value=%s%n", metaData.getKey(), metaData.getValue()); } // across this file allQueryIndices.observe(entry.getReadIndex()); totalReadLength += readLength; totalReadLengthPair += entry.getReadLengthPair(); // across all files numberOfReads++; numberOfDescriptions += entry.hasDescription() ? 1 : 0; cumulativeReadLength += readLength; if (verbose && entry.hasDescription()) { stream.println("Description found: " + entry.getDescription()); } numberOfIdentifiers += entry.hasReadIdentifier() ? 1 : 0; if (verbose && entry.hasReadIdentifier()) { stream.printf("Identifier found: %s / size=%,d%n", entry.getReadIdentifier(), readLength); } numberOfSequences += entry.hasSequence() && !entry.getSequence().isEmpty() ? 1 : 0; final boolean samplePaired = entry.hasSequencePair() && !entry.getSequencePair().isEmpty(); if (samplePaired) { numberOfSequencePairs += 1; } if (!checkedForPaired) { // Check only the very first entry. checkedForPaired = true; pairedSamples.add(samplePaired); } if (entry.hasQualityScores() && !entry.getQualityScores().isEmpty()) { numberOfQualityScores += 1; final int qualityLength = entry.getQualityScores().size(); minQualityLength = Math.min(minQualityLength, qualityLength); maxQualityLength = Math.max(maxQualityLength, qualityLength); } numberOfQualityScorePairs += entry.hasQualityScoresPair() && !entry.getQualityScoresPair().isEmpty() ? 1 : 0; // we only need to keep all the read lengths if quantiles are being computed if (computeQuantiles) { readLengths.add(readLength); } minLength = Math.min(minLength, readLength); maxLength = Math.max(maxLength, readLength); // adjust the min/max length of across all files minReadLength = Math.min(minReadLength, readLength); maxReadLength = Math.max(maxReadLength, readLength); } stream.printf("Average bytes per entry: %f%n", divide(size, allQueryIndices.count())); stream.printf("Average bytes per base: %f%n", divide(size, cumulativeReadLength)); } finally { if (reader != null) { reader.close(); } } final int numReadEntries = allQueryIndices.count(); stream.printf("Has identifiers = %s (%,d) %n", numberOfIdentifiers > 0, numberOfIdentifiers); stream.printf("Has descriptions = %s (%,d) %n", numberOfDescriptions > 0, numberOfDescriptions); stream.printf("Has sequences = %s (%,d) %n", numberOfSequences > 0, numberOfSequences); stream.printf("Has sequencePairs = %s (%,d) %n", numberOfSequencePairs > 0, numberOfSequencePairs); stream.printf("Has quality scores = %s (%,d) %n", numberOfQualityScores > 0, numberOfQualityScores); stream.printf("Has quality score Pairs = %s (%,d) %n", numberOfQualityScorePairs > 0, numberOfQualityScorePairs); stream.printf("Number of entries = %,d%n", numReadEntries); stream.printf("Min read length = %,d%n", numReadEntries > 0 ? minLength : 0); stream.printf("Max read length = %,d%n", numReadEntries > 0 ? maxLength : 0); stream.printf("Min quality length = %,d%n", numberOfQualityScores > 0 ? minQualityLength : 0); stream.printf("Max quality length = %,d%n", numberOfQualityScores > 0 ? maxQualityLength : 0); stream.printf("Avg read length = %,d%n", numReadEntries > 0 ? totalReadLength / numReadEntries : 0); stream.printf("Avg read pair length = %,d%n", numReadEntries > 0 ? totalReadLengthPair / numReadEntries : 0); // compute quantiles if (computeQuantiles) { final Percentile percentile = new Percentile(); final double[] increasingReadLengths = readLengths.toDoubleArray(); Arrays.sort(increasingReadLengths); stream.printf("Read length quantiles = [ "); for (int quantile = 1; quantile < numberOfQuantiles + 1; quantile++) { stream.printf("%,f ", percentile.evaluate(increasingReadLengths, quantile)); } stream.printf("]%n"); } }
From source file:org.drugepi.hdps.ZBiasCalculator.java
public static void scoreVariables(List<HdpsVariable> variableList) { // copy variables list List<HdpsVariable> expSortVariableList = new ArrayList<HdpsVariable>(); List<HdpsVariable> outcomeSortVariableList = new ArrayList<HdpsVariable>(); for (HdpsVariable var : variableList) { var.zBiasScore = 0; if ((var.expAssocRankingVariable != HdpsVariable.INVALID) && (var.outcomeAssocRankingVariable != HdpsVariable.INVALID)) { expSortVariableList.add(var); outcomeSortVariableList.add(var); }/*from www. j a v a 2 s . c o m*/ } // sort variables by exposure association (strongest first) Collections.sort(expSortVariableList, new HdpsVariableReverseExposureAssociationComparator()); // sort variables by outcome association (weakest first) Collections.sort(outcomeSortVariableList, new HdpsVariableReverseOutcomeAssociationComparator()); Collections.reverse(outcomeSortVariableList); // create an array of outcome strengths double[] outcomeStrengths = new double[outcomeSortVariableList.size()]; for (int i = 0; i < outcomeStrengths.length; i++) outcomeStrengths[i] = outcomeSortVariableList.get(i).outcomeAssocRankingVariable; // array that will store breaks between deciles // find the median of outcome strength Percentile pctile = new Percentile(); // Find quintiles 1 through 5 of outcome weakness // AMONG the weakest half of the variables. // List is sorted strongest first, so the weakest variables // will be at the end // quintile 1 = weakest // don't use startsOfQuintile[0] double median = pctile.evaluate(outcomeStrengths, 50.0); int searchCeiling = Arrays.binarySearch(outcomeStrengths, median); if (searchCeiling < 0) searchCeiling = -(searchCeiling + 1); int startsOfQuintile[] = new int[7]; for (int quintile = 1; quintile <= 5; quintile++) { // find the probability that *begins* this quintile double p = (quintile - 1) * 20; if (p > 0) { double quintileStartP = pctile.evaluate(outcomeStrengths, 0, searchCeiling, (quintile - 1) * 20); startsOfQuintile[quintile] = Arrays.binarySearch(outcomeStrengths, quintileStartP); if (startsOfQuintile[quintile] < 0) startsOfQuintile[quintile] = -(startsOfQuintile[quintile] + 1); } else startsOfQuintile[quintile] = 0; } startsOfQuintile[6] = searchCeiling; // score the variables, BUT make quintile 5 the weakest for (int quintile = 1; quintile <= 5; quintile++) { for (int i = startsOfQuintile[quintile]; i < startsOfQuintile[quintile + 1]; i++) { HdpsVariable v = outcomeSortVariableList.get(i); v.zBiasScore = 6 - quintile; } } // for (HdpsVariable v: outcomeSortVariableList) { // System.out.printf("%s %1.4f %d\n", v.varName, v.outcomeAssocRankingVariable, v.zBiasScore); // } }