// // FILE: Statistic.cpp // AUTHOR: Rob dot Tillaart at gmail dot com // modified at 0.3 by Gil Ross at physics dot org // VERSION: 0.4.1 // PURPOSE: Recursive statistical library for Arduino // // NOTE: 2011-01-07 Gill Ross // Rob Tillaart's Statistic library uses one-pass of the data (allowing // each value to be discarded), but expands the Sum of Squares Differences to // difference the Sum of Squares and the Average Squared. This is susceptible // to bit length precision errors with the float type (only 5 or 6 digits // absolute precision) so for long runs and high ratios of // the average value to standard deviation the estimate of the // standard error (deviation) becomes the difference of two large // numbers and will tend to zero. // // For small numbers of iterations and small Average/SE th original code is // likely to work fine. // It should also be recognised that for very large samples, questions // of stability of the sample assume greater importance than the // correctness of the asymptotic estimators. // // This recursive algorithm, which takes slightly more computation per // iteration is numerically stable. // It updates the number, mean, max, min and SumOfSquaresDiff each step to // deliver max min average, population standard error (standard deviation) and // unbiassed SE. // ------------- // // HISTORY: // 0.1 2010-10-29 initial version // 0.2 2010-10-29 stripped to minimal functionality // 0.2.01 2010-10-30 // added minimim, maximum, unbiased stdev, // changed counter to long -> int overflows @32K samples // 0.3 2011-01-07 // branched from 0.2.01 version of Rob Tillaart's code // 0.3.1 2012-11-10 minor edits // 0.3.2 2012-11-10 minor edits // changed count -> unsigned long allows for 2^32 samples // added variance() // 0.3.3 2015-03-07 // float -> double to support ARM (compiles) // moved count() sum() min() max() to .h; for optimizing compiler // 0.3.4 2017-07-31 // Refactored const in many places // [reverted] double to float on request as float is 99.99% of the cases // good enough and float(32 bit) is supported in HW for some processors. // 0.3.5 2017-09-27 // Added #include to fix uint32_t bug // 0.4.0 2020-05-13 // refactor // Added flag to switch on the use of stdDev runtime. [idea marc.recksiedl] // 0.4.1 2020-06-19 fix library.json // #include "Statistic.h" Statistic::Statistic(bool useStdDev) { clear(useStdDev); } void Statistic::clear(bool useStdDev) // useStdDev default true. { _cnt = 0; _sum = 0; _min = 0; _max = 0; _useStdDev = useStdDev; _ssqdif = 0.0; // note not _ssq but sum of square differences // which is SUM(from i = 1 to N) of f(i)-_ave_N)**2 } // adds a new value to the data-set void Statistic::add(const float value) { if (_cnt == 0) { _min = value; _max = value; } else { if (value < _min) _min = value; else if (value > _max) _max = value; } _sum += value; _cnt++; if (_useStdDev && (_cnt > 1)) { float _store = (_sum / _cnt - value); _ssqdif = _ssqdif + _cnt * _store * _store / (_cnt - 1); // ~10% faster but limits the amount of samples to 65K as _cnt*_cnt overflows // float _store = _sum - _cnt * value; // _ssqdif = _ssqdif + _store * _store / (_cnt*_cnt - _cnt); // // solution: TODO verify // _ssqdif = _ssqdif + (_store * _store / _cnt) / (_cnt - 1); } } // returns the average of the data-set added sofar float Statistic::average() const { if (_cnt == 0) return NAN; // prevent DIV0 error return _sum / _cnt; } // Population standard deviation = s = sqrt [ S ( Xi - µ )2 / N ] // http://www.suite101.com/content/how-is-standard-deviation-used-a99084 float Statistic::variance() const { if (!_useStdDev) return NAN; if (_cnt == 0) return NAN; // prevent DIV0 error return _ssqdif / _cnt; } float Statistic::pop_stdev() const { if (!_useStdDev) return NAN; if (_cnt == 0) return NAN; // prevent DIV0 error return sqrt( _ssqdif / _cnt); } float Statistic::unbiased_stdev() const { if (!_useStdDev) return NAN; if (_cnt < 2) return NAN; // prevent DIV0 error return sqrt( _ssqdif / (_cnt - 1)); } // -- END OF FILE --