下面是一个基于Groovy回答的实际Java实现。(我知道已经发布了Groovy和Kotlin实现，但对于像我这样只做Java的人来说，弄清楚如何在其他语言和Java之间转换真的很麻烦)。

(结果与他人图表相匹配)

算法实现

import java.util.ArrayList;
import java.util.Collections;
import java.util.HashMap;
import java.util.List;

import org.apache.commons.math3.stat.descriptive.SummaryStatistics;

public class SignalDetector {

    public HashMap<String, List> analyzeDataForSignals(List<Double> data, int lag, Double threshold, Double influence) {

        // init stats instance
        SummaryStatistics stats = new SummaryStatistics();

        // the results (peaks, 1 or -1) of our algorithm
        List<Integer> signals = new ArrayList<Integer>(Collections.nCopies(data.size(), 0));

        // filter out the signals (peaks) from our original list (using influence arg)
        List<Double> filteredData = new ArrayList<Double>(data);

        // the current average of the rolling window
        List<Double> avgFilter = new ArrayList<Double>(Collections.nCopies(data.size(), 0.0d));

        // the current standard deviation of the rolling window
        List<Double> stdFilter = new ArrayList<Double>(Collections.nCopies(data.size(), 0.0d));

        // init avgFilter and stdFilter
        for (int i = 0; i < lag; i++) {
            stats.addValue(data.get(i));
        }
        avgFilter.set(lag - 1, stats.getMean());
        stdFilter.set(lag - 1, Math.sqrt(stats.getPopulationVariance())); // getStandardDeviation() uses sample variance
        stats.clear();

        // loop input starting at end of rolling window
        for (int i = lag; i < data.size(); i++) {

            // if the distance between the current value and average is enough standard deviations (threshold) away
            if (Math.abs((data.get(i) - avgFilter.get(i - 1))) > threshold * stdFilter.get(i - 1)) {

                // this is a signal (i.e. peak), determine if it is a positive or negative signal
                if (data.get(i) > avgFilter.get(i - 1)) {
                    signals.set(i, 1);
                } else {
                    signals.set(i, -1);
                }

                // filter this signal out using influence
                filteredData.set(i, (influence * data.get(i)) + ((1 - influence) * filteredData.get(i - 1)));
            } else {
                // ensure this signal remains a zero
                signals.set(i, 0);
                // ensure this value is not filtered
                filteredData.set(i, data.get(i));
            }

            // update rolling average and deviation
            for (int j = i - lag; j < i; j++) {
                stats.addValue(filteredData.get(j));
            }
            avgFilter.set(i, stats.getMean());
            stdFilter.set(i, Math.sqrt(stats.getPopulationVariance()));
            stats.clear();
        }

        HashMap<String, List> returnMap = new HashMap<String, List>();
        returnMap.put("signals", signals);
        returnMap.put("filteredData", filteredData);
        returnMap.put("avgFilter", avgFilter);
        returnMap.put("stdFilter", stdFilter);

        return returnMap;

    } // end
}

主要方法

import java.text.DecimalFormat;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.HashMap;
import java.util.List;

public class Main {

    public static void main(String[] args) throws Exception {
        DecimalFormat df = new DecimalFormat("#0.000");

        ArrayList<Double> data = new ArrayList<Double>(Arrays.asList(1d, 1d, 1.1d, 1d, 0.9d, 1d, 1d, 1.1d, 1d, 0.9d, 1d,
                1.1d, 1d, 1d, 0.9d, 1d, 1d, 1.1d, 1d, 1d, 1d, 1d, 1.1d, 0.9d, 1d, 1.1d, 1d, 1d, 0.9d, 1d, 1.1d, 1d, 1d,
                1.1d, 1d, 0.8d, 0.9d, 1d, 1.2d, 0.9d, 1d, 1d, 1.1d, 1.2d, 1d, 1.5d, 1d, 3d, 2d, 5d, 3d, 2d, 1d, 1d, 1d,
                0.9d, 1d, 1d, 3d, 2.6d, 4d, 3d, 3.2d, 2d, 1d, 1d, 0.8d, 4d, 4d, 2d, 2.5d, 1d, 1d, 1d));

        SignalDetector signalDetector = new SignalDetector();
        int lag = 30;
        double threshold = 5;
        double influence = 0;

        HashMap<String, List> resultsMap = signalDetector.analyzeDataForSignals(data, lag, threshold, influence);
        // print algorithm params
        System.out.println("lag: " + lag + "\t\tthreshold: " + threshold + "\t\tinfluence: " + influence);

        System.out.println("Data size: " + data.size());
        System.out.println("Signals size: " + resultsMap.get("signals").size());

        // print data
        System.out.print("Data:\t\t");
        for (double d : data) {
            System.out.print(df.format(d) + "\t");
        }
        System.out.println();

        // print signals
        System.out.print("Signals:\t");
        List<Integer> signalsList = resultsMap.get("signals");
        for (int i : signalsList) {
            System.out.print(df.format(i) + "\t");
        }
        System.out.println();

        // print filtered data
        System.out.print("Filtered Data:\t");
        List<Double> filteredDataList = resultsMap.get("filteredData");
        for (double d : filteredDataList) {
            System.out.print(df.format(d) + "\t");
        }
        System.out.println();

        // print running average
        System.out.print("Avg Filter:\t");
        List<Double> avgFilterList = resultsMap.get("avgFilter");
        for (double d : avgFilterList) {
            System.out.print(df.format(d) + "\t");
        }
        System.out.println();

        // print running std
        System.out.print("Std filter:\t");
        List<Double> stdFilterList = resultsMap.get("stdFilter");
        for (double d : stdFilterList) {
            System.out.print(df.format(d) + "\t");
        }
        System.out.println();

        System.out.println();
        for (int i = 0; i < signalsList.size(); i++) {
            if (signalsList.get(i) != 0) {
                System.out.println("Point " + i + " gave signal " + signalsList.get(i));
            }
        }
    }
}

结果

lag: 30     threshold: 5.0      influence: 0.0
Data size: 74
Signals size: 74
Data:           1.000   1.000   1.100   1.000   0.900   1.000   1.000   1.100   1.000   0.900   1.000   1.100   1.000   1.000   0.900   1.000   1.000   1.100   1.000   1.000   1.000   1.000   1.100   0.900   1.000   1.100   1.000   1.000   0.900   1.000   1.100   1.000   1.000   1.100   1.000   0.800   0.900   1.000   1.200   0.900   1.000   1.000   1.100   1.200   1.000   1.500   1.000   3.000   2.000   5.000   3.000   2.000   1.000   1.000   1.000   0.900   1.000   1.000   3.000   2.600   4.000   3.000   3.200   2.000   1.000   1.000   0.800   4.000   4.000   2.000   2.500   1.000   1.000   1.000   
Signals:        0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   1.000   0.000   1.000   1.000   1.000   1.000   1.000   0.000   0.000   0.000   0.000   0.000   0.000   1.000   1.000   1.000   1.000   1.000   1.000   0.000   0.000   0.000   1.000   1.000   1.000   1.000   0.000   0.000   0.000   
Filtered Data:  1.000   1.000   1.100   1.000   0.900   1.000   1.000   1.100   1.000   0.900   1.000   1.100   1.000   1.000   0.900   1.000   1.000   1.100   1.000   1.000   1.000   1.000   1.100   0.900   1.000   1.100   1.000   1.000   0.900   1.000   1.100   1.000   1.000   1.100   1.000   0.800   0.900   1.000   1.200   0.900   1.000   1.000   1.100   1.200   1.000   1.000   1.000   1.000   1.000   1.000   1.000   1.000   1.000   1.000   1.000   0.900   1.000   1.000   1.000   1.000   1.000   1.000   1.000   1.000   1.000   1.000   0.800   0.800   0.800   0.800   0.800   1.000   1.000   1.000   
Avg Filter:     0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   1.003   1.003   1.007   1.007   1.003   1.007   1.010   1.003   1.000   0.997   1.003   1.003   1.003   1.000   1.003   1.010   1.013   1.013   1.013   1.010   1.010   1.010   1.010   1.010   1.007   1.010   1.010   1.003   1.003   1.003   1.007   1.007   1.003   1.003   1.003   1.000   1.000   1.007   1.003   0.997   0.983   0.980   0.973   0.973   0.970   
Std filter:     0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.000   0.060   0.060   0.063   0.063   0.060   0.063   0.060   0.071   0.073   0.071   0.080   0.080   0.080   0.077   0.080   0.087   0.085   0.085   0.085   0.083   0.083   0.083   0.083   0.083   0.081   0.079   0.079   0.080   0.080   0.080   0.077   0.077   0.075   0.075   0.075   0.073   0.073   0.063   0.071   0.080   0.078   0.083   0.089   0.089   0.086   

Point 45 gave signal 1
Point 47 gave signal 1
Point 48 gave signal 1
Point 49 gave signal 1
Point 50 gave signal 1
Point 51 gave signal 1
Point 58 gave signal 1
Point 59 gave signal 1
Point 60 gave signal 1
Point 61 gave signal 1
Point 62 gave signal 1
Point 63 gave signal 1
Point 67 gave signal 1
Point 68 gave signal 1
Point 69 gave signal 1
Point 70 gave signal 1

我在我的机器人项目中需要这样的东西。我想我可以归还Kotlin实现。

/**
* Smoothed zero-score alogrithm shamelessly copied from https://stackoverflow.com/a/22640362/6029703
* Uses a rolling mean and a rolling deviation (separate) to identify peaks in a vector
*
* @param y - The input vector to analyze
* @param lag - The lag of the moving window (i.e. how big the window is)
* @param threshold - The z-score at which the algorithm signals (i.e. how many standard deviations away from the moving mean a peak (or signal) is)
* @param influence - The influence (between 0 and 1) of new signals on the mean and standard deviation (how much a peak (or signal) should affect other values near it)
* @return - The calculated averages (avgFilter) and deviations (stdFilter), and the signals (signals)
*/
fun smoothedZScore(y: List<Double>, lag: Int, threshold: Double, influence: Double): Triple<List<Int>, List<Double>, List<Double>> {
    val stats = SummaryStatistics()
    // the results (peaks, 1 or -1) of our algorithm
    val signals = MutableList<Int>(y.size, { 0 })
    // filter out the signals (peaks) from our original list (using influence arg)
    val filteredY = ArrayList<Double>(y)
    // the current average of the rolling window
    val avgFilter = MutableList<Double>(y.size, { 0.0 })
    // the current standard deviation of the rolling window
    val stdFilter = MutableList<Double>(y.size, { 0.0 })
    // init avgFilter and stdFilter
    y.take(lag).forEach { s -> stats.addValue(s) }
    avgFilter[lag - 1] = stats.mean
    stdFilter[lag - 1] = Math.sqrt(stats.populationVariance) // getStandardDeviation() uses sample variance (not what we want)
    stats.clear()
    //loop input starting at end of rolling window
    (lag..y.size - 1).forEach { i ->
        //if the distance between the current value and average is enough standard deviations (threshold) away
        if (Math.abs(y[i] - avgFilter[i - 1]) > threshold * stdFilter[i - 1]) {
            //this is a signal (i.e. peak), determine if it is a positive or negative signal
            signals[i] = if (y[i] > avgFilter[i - 1]) 1 else -1
            //filter this signal out using influence
            filteredY[i] = (influence * y[i]) + ((1 - influence) * filteredY[i - 1])
        } else {
            //ensure this signal remains a zero
            signals[i] = 0
            //ensure this value is not filtered
            filteredY[i] = y[i]
        }
        //update rolling average and deviation
        (i - lag..i - 1).forEach { stats.addValue(filteredY[it]) }
        avgFilter[i] = stats.getMean()
        stdFilter[i] = Math.sqrt(stats.getPopulationVariance()) //getStandardDeviation() uses sample variance (not what we want)
        stats.clear()
    }
    return Triple(signals, avgFilter, stdFilter)
}

带有验证图的示例项目可以在github上找到。

这是一个Python实现的鲁棒峰值检测算法算法。

初始化和计算部分被分开，只有filtered_y数组被保留，它的最大大小等于延迟，因此内存没有增加。(结果与上述答案相同)。 为了绘制图形，还保留了标签数组。

我做了一个github要点。

import numpy as np
import pylab

def init(x, lag, threshold, influence):
    '''
    Smoothed z-score algorithm
    Implementation of algorithm from https://stackoverflow.com/a/22640362/6029703
    '''

    labels = np.zeros(lag)
    filtered_y = np.array(x[0:lag])
    avg_filter = np.zeros(lag)
    std_filter = np.zeros(lag)
    var_filter = np.zeros(lag)

    avg_filter[lag - 1] = np.mean(x[0:lag])
    std_filter[lag - 1] = np.std(x[0:lag])
    var_filter[lag - 1] = np.var(x[0:lag])

    return dict(avg=avg_filter[lag - 1], var=var_filter[lag - 1],
                std=std_filter[lag - 1], filtered_y=filtered_y,
                labels=labels)


def add(result, single_value, lag, threshold, influence):
    previous_avg = result['avg']
    previous_var = result['var']
    previous_std = result['std']
    filtered_y = result['filtered_y']
    labels = result['labels']

    if abs(single_value - previous_avg) > threshold * previous_std:
        if single_value > previous_avg:
            labels = np.append(labels, 1)
        else:
            labels = np.append(labels, -1)

        # calculate the new filtered element using the influence factor
        filtered_y = np.append(filtered_y, influence * single_value
                               + (1 - influence) * filtered_y[-1])
    else:
        labels = np.append(labels, 0)
        filtered_y = np.append(filtered_y, single_value)

    # update avg as sum of the previuos avg + the lag * (the new calculated item - calculated item at position (i - lag))
    current_avg_filter = previous_avg + 1. / lag * (filtered_y[-1]
            - filtered_y[len(filtered_y) - lag - 1])

    # update variance as the previuos element variance + 1 / lag * new recalculated item - the previous avg -
    current_var_filter = previous_var + 1. / lag * ((filtered_y[-1]
            - previous_avg) ** 2 - (filtered_y[len(filtered_y) - 1
            - lag] - previous_avg) ** 2 - (filtered_y[-1]
            - filtered_y[len(filtered_y) - 1 - lag]) ** 2 / lag)  # the recalculated element at pos (lag) - avg of the previuos - new recalculated element - recalculated element at lag pos ....

    # calculate standard deviation for current element as sqrt (current variance)
    current_std_filter = np.sqrt(current_var_filter)

    return dict(avg=current_avg_filter, var=current_var_filter,
                std=current_std_filter, filtered_y=filtered_y[1:],
                labels=labels)

lag = 30
threshold = 5
influence = 0

y = np.array([1,1,1.1,1,0.9,1,1,1.1,1,0.9,1,1.1,1,1,0.9,1,1,1.1,1,1,1,1,1.1,0.9,1,1.1,1,1,0.9,
       1,1.1,1,1,1.1,1,0.8,0.9,1,1.2,0.9,1,1,1.1,1.2,1,1.5,1,3,2,5,3,2,1,1,1,0.9,1,1,3,
       2.6,4,3,3.2,2,1,1,0.8,4,4,2,2.5,1,1,1])

# Run algo with settings from above
result = init(y[:lag], lag=lag, threshold=threshold, influence=influence)

i = open('quartz2', 'r')
for i in y[lag:]:
    result = add(result, i, lag, threshold, influence)

# Plot result
pylab.subplot(211)
pylab.plot(np.arange(1, len(y) + 1), y)
pylab.subplot(212)
pylab.step(np.arange(1, len(y) + 1), result['labels'], color='red',
           lw=2)
pylab.ylim(-1.5, 1.5)
pylab.show()

下面是ZSCORE算法的PHP实现:

<?php
$y = array(1,7,1.1,1,0.9,1,1,1.1,1,0.9,1,1.1,1,1,0.9,1,1,1.1,1,1,1,1,1.1,0.9,1,1.1,1,1,0.9,
       1,1.1,1,1,1.1,1,0.8,0.9,1,1.2,0.9,1,1,1.1,1.2,1,1.5,10,3,2,5,3,2,1,1,1,0.9,1,1,3,
       2.6,4,3,3.2,2,1,1,0.8,4,4,2,2.5,1,1,1);

function mean($data, $start, $len) {
    $avg = 0;
    for ($i = $start; $i < $start+ $len; $i ++)
        $avg += $data[$i];
    return $avg / $len;
}
    
function stddev($data, $start,$len) {
    $mean = mean($data,$start,$len);
    $dev = 0;
    for ($i = $start; $i < $start+$len; $i++) 
        $dev += (($data[$i] - $mean) * ($data[$i] - $mean));
    return sqrt($dev / $len);
}

function zscore($data, $len, $lag= 20, $threshold = 1, $influence = 1) {

    $signals = array();
    $avgFilter = array();
    $stdFilter = array();
    $filteredY = array();
    $avgFilter[$lag - 1] = mean($data, 0, $lag);
    $stdFilter[$lag - 1] = stddev($data, 0, $lag);
    
    for ($i = 0; $i < $len; $i++) {
        $filteredY[$i] = $data[$i];
        $signals[$i] = 0;
    }


    for ($i=$lag; $i < $len; $i++) {
        if (abs($data[$i] - $avgFilter[$i-1]) > $threshold * $stdFilter[$lag - 1]) {
            if ($data[$i] > $avgFilter[$i-1]) {
                $signals[$i] = 1;
            }
            else {
                $signals[$i] = -1;
            }
            $filteredY[$i] = $influence * $data[$i] + (1 - $influence) * $filteredY[$i-1];
        } 
        else {
            $signals[$i] = 0;
            $filteredY[$i] = $data[$i];
        }
        
        $avgFilter[$i] = mean($filteredY, $i - $lag, $lag);
        $stdFilter[$i] = stddev($filteredY, $i - $lag, $lag);
    }
    return $signals;
}

$sig = zscore($y, count($y));

print_r($y); echo "<br><br>";
print_r($sig); echo "<br><br>";

for ($i = 0; $i < count($y); $i++) echo $i. " " . $y[$i]. " ". $sig[$i]."<br>";

如果你的数据在一个数据库表中，这里是一个简单的z-score算法的SQL版本:

with data_with_zscore as (
    select
        date_time,
        value,
        value / (avg(value) over ()) as pct_of_mean,
        (value - avg(value) over ()) / (stdev(value) over ()) as z_score
    from {{tablename}}  where datetime > '2018-11-26' and datetime < '2018-12-03'
)


-- select all
select * from data_with_zscore 

-- select only points greater than a certain threshold
select * from data_with_zscore where z_score > abs(2)

我们尝试在我们的数据集上使用平滑的z-score算法，这导致了过度敏感或不敏感(取决于参数如何调整)，几乎没有中间地带。在我们站点的交通信号中，我们观察到一个低频基线，它代表了每天的周期，即使有最好的可能参数(如下所示)，它仍然在第4天下降，特别是因为大多数数据点被认为是异常的。

在原始z-score算法的基础上，我们提出了一种通过反向滤波来解决这个问题的方法。改进后的算法及其在电视商业流量归因中的应用详见我们的团队博客。