igg2/CumulStats

A cumulative statistics class for any number of samples using Welford's algorithm

CumulStats

A C++ class for cumulative statistics for any number of samples using Welford's algorithm

// Accumulate statistics without storing any samples
CumulStats randSamples;
uint32_t i;

for (i = 0; i < 10000; i++) {
	// Generate a random sample and add it to the cumulative statistics
	long sample = random(0, 9999);
	randSamples.add(sample/100.0);

	// Statistics can be retrieved at any time during sampling
	if (i && i % 100 == 0) {
		// Number of samples so far and cumulative average
		Serial.println (randSamples.n());
		Serial.println (randSamples.mean());

		// Minimum & maximum
		Serial.println (randSamples.minimum());
		Serial.println (randSamples.maximum());

		// Variance and standard deviation (unbiased, N-1)
		// N.B.: These require more than two samples (otherwise return 0.0)
		Serial.println (randSamples.var());
		Serial.println (randSamples.std());
		Serial.println ();
	}

	// Reset the stats and keep going
	if (i && i % 1000 == 0) {
		randSamples.reset();
		Serial.println ();
		Serial.println ();
	}
}