PDA

View Full Version : Accuracy and precision



Brian Deakin
04-17-2016, 3:55 AM
Accuracy and precision

From Wikipedia, the free encyclopedi
Precision is a description of random errors (https://en.wikipedia.org/wiki/Random_errors), a measure of statistical variability (https://en.wikipedia.org/wiki/Statistical_variability).
Accuracy has two definitions:


more commonly, it is a description of systematic errors (https://en.wikipedia.org/wiki/Systematic_errors), a measure of statistical bias (https://en.wikipedia.org/wiki/Statistical_bias);
alternatively, the ISO (https://en.wikipedia.org/wiki/International_Organization_for_Standardization) defines accuracy as describing both types of observational error (https://en.wikipedia.org/wiki/Observational_error) above (preferring the term trueness for the common definition of accuracy).


Contents




Common definition


https://upload.wikimedia.org/wikipedia/commons/thumb/3/38/Accuracy_and_precision.svg/300px-Accuracy_and_precision.svg.png (https://en.wikipedia.org/wiki/File:Accuracy_and_precision.svg)

Accuracy is the proximity of measurement results to the true value; precision, the repeatability, or reproducibility of the measurement


In the fields of science (https://en.wikipedia.org/wiki/Science), engineering (https://en.wikipedia.org/wiki/Engineering) and statistics (https://en.wikipedia.org/wiki/Statistics), the accuracy of a measurement (https://en.wikipedia.org/wiki/Measurement) system is the degree of closeness of measurements of a quantity (https://en.wikipedia.org/wiki/Quantity) to that quantity's true value (https://en.wikipedia.org/wiki/Value_(mathematics)).[1] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-metrology_terms-1) The precision of a measurement system, related to reproducibility (https://en.wikipedia.org/wiki/Reproducibility)and repeatability (https://en.wikipedia.org/wiki/Repeatability), is the degree to which repeated measurements under unchanged conditions show the same results (https://en.wikipedia.org/wiki/Result).[1] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-metrology_terms-1)[2] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-Taylor-2)Although the two words precision and accuracy can be synonymous (https://en.wikipedia.org/wiki/Synonymous) in colloquial (https://en.wikipedia.org/wiki/Colloquial) use, they are deliberately contrasted in the context of the scientific method (https://en.wikipedia.org/wiki/Scientific_method).
A measurement system can be accurate but not precise, precise but not accurate, neither, or both. For example, if an experiment contains a systematic error (https://en.wikipedia.org/wiki/Systematic_error), then increasing the sample size (https://en.wikipedia.org/wiki/Sample_size) generally increases precision but does not improve accuracy. The result would be a consistent yet inaccurate string of results from the flawed experiment. Eliminating the systematic error improves accuracy but does not change precision.
A measurement system is considered valid if it is both accurate and precise. Related terms include bias (non-random (https://en.wikipedia.org/wiki/Random) or directed effects caused by a factor or factors unrelated to the independent variable (https://en.wikipedia.org/wiki/Independent_variable)) and error (random variability).
The terminology is also applied to indirect measurements—that is, values obtained by a computational procedure from observed data.
In addition to accuracy and precision, measurements may also have a measurement resolution (https://en.wikipedia.org/wiki/Sensor#Resolution), which is the smallest change in the underlying physical quantity that produces a response in the measurement.
In numerical analysis (https://en.wikipedia.org/wiki/Numerical_analysis), accuracy is also the nearness of a calculation to the true value; while precision is the resolution of the representation, typically defined by the number of decimal or binary digits.
Statistical literature prefers to use the terms bias (https://en.wikipedia.org/wiki/Bias_(statistics)) and variability (https://en.wikipedia.org/wiki/Variability_(statistics)) instead of accuracy and precision: bias is the amount of inaccuracy and variability is the amount of imprecision.
Quantification[edit (https://en.wikipedia.org/w/index.php?title=Accuracy_and_precision&action=edit&section=2)]

See also: False precision (https://en.wikipedia.org/wiki/False_precision)
In industrial instrumentation, accuracy is the measurement tolerance, or transmission of the instrument and defines the limits of the errors made when the instrument is used in normal operating conditions.[3] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-3)
Ideally a measurement device is both accurate and precise, with measurements all close to and tightly clustered around the true value. The accuracy and precision of a measurement process is usually established by repeatedly measuring some traceable (https://en.wikipedia.org/wiki/Traceability) reference standard (https://en.wikipedia.org/wiki/Technical_standard). Such standards are defined in the International System of Units (https://en.wikipedia.org/wiki/SI)(abbreviated SI from French: Système international d'unités) and maintained by national standards organizations (https://en.wikipedia.org/wiki/Standards_organization) such as the National Institute of Standards and Technology (https://en.wikipedia.org/wiki/National_Institute_of_Standards_and_Technology) in the United States.
This also applies when measurements are repeated and averaged. In that case, the term standard error (https://en.wikipedia.org/wiki/Standard_error_(statistics)) is properly applied: the precision of the average is equal to the known standard deviation of the process divided by the square root of the number of measurements averaged. Further, the central limit theorem (https://en.wikipedia.org/wiki/Central_limit_theorem) shows that the probability distribution (https://en.wikipedia.org/wiki/Probability_distribution) of the averaged measurements will be closer to a normal distribution than that of individual measurements.
With regard to accuracy we can distinguish:


the difference between the mean (https://en.wikipedia.org/wiki/Mean) of the measurements and the reference value, the bias (https://en.wikipedia.org/wiki/Bias_of_an_estimator). Establishing and correcting for bias is necessary for calibration (https://en.wikipedia.org/wiki/Calibration).
the combined effect of that and precision.

A common convention in science and engineering is to express accuracy and/or precision implicitly by means of significant figures (https://en.wikipedia.org/wiki/Significant_figures). Here, when not explicitly stated, the margin of error is understood to be one-half the value of the last significant place. For instance, a recording of 843.6 m, or 843.0 m, or 800.0 m would imply a margin of 0.05 m (the last significant place is the tenths place), while a recording of 8,436 m would imply a margin of error of 0.5 m (the last significant digits are the units).
A reading of 8,000 m, with trailing zeroes and no decimal point, is ambiguous; the trailing zeroes may or may not be intended as significant figures. To avoid this ambiguity, the number could be represented in scientific notation: 8.0 × 103 m indicates that the first zero is significant (hence a margin of 50 m) while 8.000 × 103 m indicates that all three zeroes are significant, giving a margin of 0.5 m. Similarly, it is possible to use a multiple of the basic measurement unit: 8.0 km is equivalent to 8.0 × 103 m. In fact, it indicates a margin of 0.05 km (50 m). However, reliance on this convention can lead to false precision (https://en.wikipedia.org/wiki/False_precision) errors when accepting data from sources that do not obey it.
Precision is sometimes stratified into:


Repeatability — the variation arising when all efforts are made to keep conditions constant by using the same instrument and operator, and repeating during a short time period; and
Reproducibility — the variation arising using the same measurement process among different instruments and operators, and over longer time periods.

ISO Definition (ISO 5725)[edit (https://en.wikipedia.org/w/index.php?title=Accuracy_and_precision&action=edit&section=3)]


https://upload.wikimedia.org/wikipedia/commons/thumb/9/92/Accuracy_%28trueness_and_precision%29.svg/300px-Accuracy_%28trueness_and_precision%29.svg.png (https://en.wikipedia.org/wiki/File:Accuracy_(trueness_and_precision).svg)

According to ISO 5725-1, Accuracy consists of Trueness (proximity of measurement results to the true value) and Precision (repeatability or reproducibility of the measurement)


A shift in the meaning of these terms appeared with the publication of the ISO 5725 series of standards in 1994, which is also reflected in the 2008 issue of the "BIPM International Vocabulary of Metrology" (VIM), items 2.13 and 2.14.[1] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-metrology_terms-1)
According to ISO 5725-1,[4] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-4) the general term "accuracy" is used to describe the closeness of a measurement to the true value. When the term is applied to sets of measurements of the same measurand, it involves a component of random error and a component of systematic error. In this case trueness is the closeness of the mean of a set of measurement results to the actual (true) value and precision is the closeness of agreement among a set of results.
ISO 5725-1 and VIM also avoid the use of the term "bias (https://en.wikipedia.org/wiki/Bias_(statistics))", previously specified in BS 5497-1,[5] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-5) because it has different connotations outside the fields of science and engineering, as in medicine and law.


Accuracy of a target grouping (https://en.wikipedia.org/wiki/Grouping_(firearms)) according to BIPM and ISO 5725




https://upload.wikimedia.org/wikipedia/commons/thumb/1/10/High_accuracy_Low_precision.svg/150px-High_accuracy_Low_precision.svg.png (https://en.wikipedia.org/wiki/File:High_accuracy_Low_precision.svg)


Low accuracy, poor precision, good trueness




https://upload.wikimedia.org/wikipedia/commons/thumb/3/3a/High_precision_Low_accuracy.svg/150px-High_precision_Low_accuracy.svg.png (https://en.wikipedia.org/wiki/File:High_precision_Low_accuracy.svg)


Low accuracy, good precision, poor trueness





In binary classification[edit (https://en.wikipedia.org/w/index.php?title=Accuracy_and_precision&action=edit&section=4)]

Main article: Evaluation of binary classifiers (https://en.wikipedia.org/wiki/Evaluation_of_binary_classifiers)


https://upload.wikimedia.org/wikipedia/en/thumb/f/f2/Edit-clear.svg/20px-Edit-clear.svg.png
This section is too long (https://en.wikipedia.org/wiki/Wikipedia:Article_size). Considersplitting (https://en.wikipedia.org/wiki/Wikipedia:Splitting) it into new pages, addingsubheadings (https://en.wikipedia.org/wiki/Help:Section#Subsections), or condensing (https://en.wikipedia.org/wiki/Wikipedia:Summary_style) it.(September 2014)








Accuracy is also used as a statistical measure of how well a binary classification (https://en.wikipedia.org/wiki/Binary_classification) test correctly identifies or excludes a condition.
That is, the accuracy is the proportion of true results (both true positives (https://en.wikipedia.org/wiki/True_positive) and true negatives (https://en.wikipedia.org/wiki/True_negative)) among the total number of cases examined.[6] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-6) To make the context clear by the semantics, it is often referred to as the "Rand accuracy" or "Rand index (https://en.wikipedia.org/wiki/Rand_index)".[7] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-7)[8] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-8)[9] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-9) It is a parameter of the test.
https://upload.wikimedia.org/math/8/5/f/85fb106488e3cb8c02e397c917222ad4.pngAn accuracy of 100% means that the measured values are exactly the same as the given values.
On the other hand, precision or positive predictive value (https://en.wikipedia.org/wiki/Positive_predictive_value) is defined as the proportion of the true positives against all the positive results (both true positives and false positives (https://en.wikipedia.org/wiki/False_positive))
https://upload.wikimedia.org/math/0/5/c/05c24c4d7ffa993d678081d8243d9ac8.pngAlso see Sensitivity and specificity (https://en.wikipedia.org/wiki/Sensitivity_and_specificity).
Accuracy may be determined from sensitivity and specificity, provided prevalence (https://en.wikipedia.org/wiki/Prevalence) is known, using the equation:
https://upload.wikimedia.org/math/0/5/7/05796b84fc84aeeb8260a838ceac684f.pngThe accuracy paradox (https://en.wikipedia.org/wiki/Accuracy_paradox) for predictive analytics (https://en.wikipedia.org/wiki/Predictive_analytics) states that predictive models with a given level of accuracy may have greater predictive power (https://en.wikipedia.org/wiki/Predictive_power) than models with higher accuracy. It may be better to avoid the accuracy metric in favor of other metrics such as precision and recall (https://en.wikipedia.org/wiki/Precision_and_recall).[citation needed (https://en.wikipedia.org/wiki/Wikipedia:Citation_needed)] In situations where the minority class is more important, F-measure (https://en.wikipedia.org/wiki/F-measure)may be more appropriate, especially in situations with very skewed class imbalance.
Another useful performance measure is the balanced accuracy[10] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-Brodersen2010-10) which avoids inflated performance estimates on imbalanced datasets. It is defined as the arithmetic mean of sensitivity and specificity, or the average accuracy obtained on either class:
https://upload.wikimedia.org/math/e/2/0/e200da5e436cda83a8cab14fcb3a275d.pnghttps://upload.wikimedia.org/math/7/c/d/7cdf4b6c2980d6b6d09d7f5a46a95de2.pngIf the classifier performs equally well on either class, this term reduces to the conventional accuracy (i.e., the number of correct predictions divided by the total number of predictions). In contrast, if the conventional accuracy is above chance only because the classifier takes advantage of an imbalanced test set, then the balanced accuracy, as appropriate, will drop to chance.[10] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-Brodersen2010-10) A closely related chance corrected measure is:
https://upload.wikimedia.org/math/2/7/b/27bf7aacdb946726d9c17e34b6838458.png[11] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-Powers2011-11)A direct approach to debiasing and renormalizing Accuracy is Cohen's kappa (https://en.wikipedia.org/wiki/Cohen%27s_kappa), whilst Informedness has been shown to be a Kappa-family debiased renormalization of Recall.[12] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-12)Informedness and Kappa have the advantage that chance level is defined to be 0, and they have the form of a probability. Informedness has the stronger property that it is the probability that an informed decision is made (rather than a guess), when positive. When negative this is still true for the absolutely value of Informedness, but the information has been used to force an incorrect response.[11] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-Powers2011-11)
In psychometrics and psychophysics[edit (https://en.wikipedia.org/w/index.php?title=Accuracy_and_precision&action=edit&section=5)]

In psychometrics (https://en.wikipedia.org/wiki/Psychometrics) and psychophysics (https://en.wikipedia.org/wiki/Psychophysics), the term accuracy is interchangeably used with validity (https://en.wikipedia.org/wiki/Validity_(statistics)) and constant error. Precision is a synonym for reliability (https://en.wikipedia.org/wiki/Reliability_(statistics)) and variable error. The validity of a measurement instrument or psychological test is established through experiment or correlation with behavior. Reliability is established with a variety of statistical techniques, classically through an internal consistency test like Cronbach's alpha (https://en.wikipedia.org/wiki/Cronbach%27s_alpha) to ensure sets of related questions have related responses, and then comparison of those related question between reference and target population.[citation needed (https://en.wikipedia.org/wiki/Wikipedia:Citation_needed)]
In logic simulation[edit (https://en.wikipedia.org/w/index.php?title=Accuracy_and_precision&action=edit&section=6)]

In logic simulation (https://en.wikipedia.org/wiki/Logic_simulation), a common mistake in evaluation of accurate models is to compare a logic simulation model (https://en.wikipedia.org/wiki/Electronic_circuit_simulation) to a transistor (https://en.wikipedia.org/wiki/Transistor) circuit simulation model (https://en.wikipedia.org/wiki/Transistor_models). This is a comparison of differences in precision, not accuracy. Precision is measured with respect to detail and accuracy is measured with respect to reality.[13] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-13)[14] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-14)
In information systems[edit (https://en.wikipedia.org/w/index.php?title=Accuracy_and_precision&action=edit&section=7)]



https://upload.wikimedia.org/wikipedia/en/thumb/f/f2/Edit-clear.svg/20px-Edit-clear.svg.png
This section may be confusing or unclear (https://en.wikipedia.org/wiki/Wikipedia:Vagueness) to readers. (March 2013)


The concepts of accuracy and precision have also been studied in the context of databases, information systems and their sociotechnical context. The necessary extension of these two concepts on the basis of theory of science suggests that they (as well as data quality (https://en.wikipedia.org/wiki/Data_quality) and information quality (https://en.wikipedia.org/wiki/Information_quality)) should be centered on accuracy defined as the closeness to the true value seen as the degree of agreement of readings or of calculated values of one same conceived entity, measured or calculated by different methods, in the context of maximum possible disagreement.[15] (https://en.wikipedia.org/wiki/Accuracy_and_precision#cite_note-15)
Further information: Precision and recall (https://en.wikipedia.org/wiki/Precision_and_recall)



https://upload.wikimedia.org/wikipedia/commons/thumb/f/f8/Wiktionary-logo-en.svg/37px-Wiktionary-logo-en.svg.png

Matt Day
04-17-2016, 7:44 AM
I'm scratching my head....

Pat Barry
04-17-2016, 8:00 AM
I'm good with the common definition.

roger wiegand
04-17-2016, 8:04 AM
So you're suggesting using stop blocks instead of a ruler to achieve precision, if not accuracy?

Paolo Trevisan
04-17-2016, 9:18 AM
point being
...

daryl moses
04-17-2016, 9:33 AM
About eight more clicks to the right and you'll be in the bullseye:rolleyes:

Jebediah Eckert
04-17-2016, 9:33 AM
I have often been the victim of "false precision".:confused:

Michael Weber
04-17-2016, 10:41 AM
As a former untrained metrologist in an industrial R&D lab I could seldom make technicians and sometimes engineers understand the difference between accuracy and resolution.

Tom Stenzel
04-17-2016, 10:47 AM
I think all this translates to:

If you can't hit anything when hunting small game, hunt large game.*

-Tom

*Slight change of something that Jed Clampett said on the Beverly Hillbillies

Frederick Skelly
04-17-2016, 11:31 AM
point being...

Yeah Brian, I think maybe the point you were making got deleted accidentally - without it, this sorta came out of nowhere. Did some other post trigger this thought? What were you driving at Sir?

Bert Kemp
04-17-2016, 1:20 PM
Was there a question or a point to be made here:confused:

Greg Peterson
04-17-2016, 1:45 PM
I have to admit I never gave any thought to the definition of accuracy or precision.
Even though I spent a lot of time making my machines accurate, yet this does garuantee precision. Precision is where the skill lies.
As Lee Trevino said, "It ain't the arrow, it's the Indian."

Allan Speers
04-17-2016, 3:43 PM
I am incredibly precise in my inaccuracy. - I make the same mistakes, in exactly the same way, over & over & over..... :o

Kev Williams
04-17-2016, 5:14 PM
Accuracy and precision don't matter in 'real life'...

to wit:
I have a half dozen or so air pressure gauges. No 2 give me the same reading... have no idea which one, if any, is correct?

Same with the kitchen, I have 2 identical digital meat thermometers, and a third spring-dial type. Had all 3 in the same ham the other day, temps were 152, 155, and 157...

how can anything be accurate when the measuring devices aren't?

;)

Frederick Skelly
04-17-2016, 5:31 PM
Accuracy and precision don't matter in 'real life'...

to wit:
I have a half dozen or so air pressure gauges. No 2 give me the same reading... have no idea which one, if any, is correct?

Same with the kitchen, I have 2 identical digital meat thermometers, and a third spring-dial type. Had all 3 in the same ham the other day, temps were 152, 155, and 157...

how can anything be accurate when the measuring devices aren't?

;)

Reminds me of that old saying "A man with one clock always knows what time it is. A man with two clocks never does."

Howard Garner
04-17-2016, 8:05 PM
Accuracy and precision don't matter in 'real life'...

how can anything be accurate when the measuring devices aren't?

;)

That's why in the Navy we had to have everything calibrated periodically.
If an item was out by just a day, we were not allowed to use it.
Such was the life working on nuclear missile systems.

Howard Garner MTC

Myk Rian
04-17-2016, 8:58 PM
https://upload.wikimedia.org/wikipedia/commons/thumb/1/10/High_accuracy_Low_precision.svg/150px-High_accuracy_Low_precision.svg.png
(https://en.wikipedia.org/wiki/File:High_accuracy_Low_precision.svg)



My target at 15 yards. 9mm Glock 26.





https://upload.wikimedia.org/wikipedia/commons/thumb/3/3a/High_precision_Low_accuracy.svg/150px-High_precision_Low_accuracy.svg.png (https://en.wikipedia.org/wiki/File:High_precision_Low_accuracy.svg)


Using too much of the end of your finger on the trigger.
Anticipating recoil.

Tom Stenzel
04-17-2016, 10:07 PM
My target before shooting at it with my muzzle loader:

335930

My target after shooting at with my muzzle loader:

335930

My shooting, if nothing else, is consistent. Isn't consistency a form of precision?

:rolleyes:

-Tom