Java Code Examples for java.util.DoubleSummaryStatistics#getAverage()

public SummaryStatistics summarizeTransactions() {

        final DoubleSummaryStatistics doubleSummaryStatistics = bankTransactions.stream()
                .mapToDouble(BankTransaction::getAmount)
                .summaryStatistics();

        return new SummaryStatistics(doubleSummaryStatistics.getSum(),
                                     doubleSummaryStatistics.getMax(),
                                     doubleSummaryStatistics.getMin(),
                                     doubleSummaryStatistics.getAverage());
    }
 

/**
 * Return a new value object of the statistical summary, currently
 * represented by the {@code statistics} object.
 *
 * @param statistics the creating (mutable) statistics class
 * @return the statistical moments
 */
public static DoubleSummary of(final DoubleSummaryStatistics statistics) {
	return new DoubleSummary(
		statistics.getCount(),
		statistics.getMin(),
		statistics.getMax(),
		statistics.getSum(),
		statistics.getAverage()
	);
}
 

public static void main(String[] args) {
	if (args.length < 2 || !new File(args[1]).isFile()) {
		System.err.println("Please provide a train file/directory and a test file for this example");
		return;
	}
	File train = new File(args[0]);
	File test = new File(args[1]);

	Lexer lexer = new JavaLexer();  // Use a Java lexer; if your code is already lexed, use whitespace or tokenized lexer
	LexerRunner lexerRunner = new LexerRunner(lexer, false);  // Don't model lines in isolation for code files.
															  // We will still get per-line, per-token entropies
	lexerRunner.setSentenceMarkers(true);  // Add start and end markers to the files
	lexerRunner.setExtension("java");  // We only lex Java files

	Vocabulary vocabulary = new Vocabulary();  // Create an empty vocabulary
	
	Model model = new JMModel(6, new GigaCounter());  // Standard smoothing for code, giga-counter for large corpora
	model = MixModel.standard(model, new CacheModel());  // Use a simple cache model; see JavaRunner for more options
	ModelRunner modelRunner = new ModelRunner(model, lexerRunner, vocabulary); // Use above lexer and vocabulary
	modelRunner.learnDirectory(train);  // Teach the model all the data in "train"
	
	// Modeling one file gives us entropy per-line, per token in nested list. See also modelRunner.modelDirectory
	List<List<Double>> fileEntropies = modelRunner.modelFile(test);
	List<List<String>> fileTokens = lexerRunner.lexFile(test)  // Let's also retrieve the tokens on each line
			.map(l -> l.collect(Collectors.toList()))
			.collect(Collectors.toList());
	for (int i = 0; i < fileEntropies.size(); i++) {
		List<String> lineTokens = fileTokens.get(i);
		List<Double> lineEntropies = fileEntropies.get(i);
		
		// First use Java's stream API to summarize entropies on this line
		// (see modelRunner.getStats for summarizing file or directory results)
		DoubleSummaryStatistics lineStatistics = lineEntropies.stream()
				.mapToDouble(Double::doubleValue)
				.summaryStatistics();
		double averageEntropy = lineStatistics.getAverage();
		
		// Then, print out the average entropy and the entropy for every token on this line
		System.out.printf("Line %d, avg.: %.4f, tokens:", i + 1, averageEntropy);
		for (int j = 0; j < lineTokens.size(); j++) {
			System.out.printf(" %s: %.4f", lineTokens.get(j), lineEntropies.get(j));
		}
		System.out.println();
	}
}