Outlier
Outlier
In most larger samplings of data, some data points will be further away from the sample mean than what is
deemed reasonable. This can be due to incidental systematic error or flaws in the theory that generated an
assumed family of probability distributions, or it may be that some observations are far from the center of
the data. Outlier points can therefore indicate faulty data, erroneous procedures, or areas where a certain
theory might not be valid. However, in large samples, a small number of outliers is to be expected (and not
due to any anomalous condition).
Outliers, being the most extreme observations, may include the sample maximum or sample minimum, or
both, depending on whether they are extremely high or low. However, the sample maximum and minimum
are not always outliers because they may not be unusually far from other observations.
Naive interpretation of statistics derived from data sets that include outliers may be misleading. For
example, if one is calculating the average temperature of 10 objects in a room, and nine of them are
between 20 and 25 degrees Celsius, but an oven is at 175 °C, the median of the data will be between 20
and 25 °C but the mean temperature will be between 35.5 and 40 °C. In this case, the median better reflects
the temperature of a randomly sampled object (but not the temperature in the room) than the mean; naively
interpreting the mean as "a typical sample", equivalent to the median, is incorrect. As illustrated in this case,
outliers may indicate data points that belong to a different population than the rest of the sample set.
Estimators capable of coping with outliers are said to be robust: the median is a robust statistic of central
tendency, while the mean is not.[5] However, the mean is generally a more precise estimator.[6]
In general, if the nature of the population distribution is known a priori, it is possible to test if the number of
outliers deviate significantly from what can be expected: for a given cutoff (so samples fall beyond the
cutoff with probability p) of a given distribution, the number of outliers will follow a binomial distribution
with parameter p, which can generally be well-approximated by the Poisson distribution with λ = pn. Thus
if one takes a normal distribution with cutoff 3 standard deviations from the mean, p is approximately 0.3%,
and thus for 1000 trials one can approximate the number of samples whose deviation exceeds 3 sigmas by a
Poisson distribution with λ = 3.
Causes
Outliers can have many anomalous causes. A physical apparatus for taking measurements may have
suffered a transient malfunction. There may have been an error in data transmission or transcription.
Outliers arise due to changes in system behaviour, fraudulent behaviour, human error, instrument error or
simply through natural deviations in populations. A sample may have been contaminated with elements
from outside the population being examined. Alternatively, an outlier could be the result of a flaw in the
assumed theory, calling for further investigation by the researcher. Additionally, the pathological appearance
of outliers of a certain form appears in a variety of datasets, indicating that the causative mechanism for the
data might differ at the extreme end (King effect).
Model-based methods which are commonly used for identification assume that the data are from a normal
distribution, and identify observations which are deemed "unlikely" based on mean and standard deviation:
Chauvenet's criterion
Grubbs's test for outliers
Dixon's Q test
ASTM E178: Standard Practice for Dealing With Outlying Observations[14]
Mahalanobis distance and leverage are often used to detect outliers, especially in the
development of linear regression models.
Subspace and correlation based techniques for high-dimensional numerical data[13]
Peirce's criterion
It is proposed to determine in a series of observations the limit of error, beyond which all
observations involving so great an error may be rejected, provided there are as many as such
observations. The principle upon which it is proposed to solve this problem is, that the
proposed observations should be rejected when the probability of the system of errors obtained
by retaining them is less than that of the system of errors obtained by their rejection multiplied
by the probability of making so many, and no more, abnormal observations. (Quoted in the
editorial note on page 516 to Peirce (1982 edition) from A Manual of Astronomy 2:558 by
Chauvenet.) [15][16][17][18]
Tukey's fences
Other methods flag observations based on measures such as the interquartile range. For example, if and
are the lower and upper quartiles respectively, then one could define an outlier to be any observation
outside the range:
for some nonnegative constant . John Tukey proposed this test, where indicates an "outlier", and
indicates data that is "far out".[19]
In anomaly detection
In various domains such as, but not limited to, statistics, signal processing, finance, econometrics,
manufacturing, networking and data mining, the task of anomaly detection may take other approaches.
Some of these may be distance-based[20][21] and density-based such as Local Outlier Factor (LOF).[22]
Some approaches may use the distance to the k-nearest neighbors to label observations as outliers or non-
outliers.[23]
The modified Thompson Tau test is a method used to determine if an outlier exists in a data set. The
strength of this method lies in the fact that it takes into account a data set's standard deviation, average and
provides a statistically determined rejection zone; thus providing an objective method to determine if a data
point is an outlier.[24] How it works: First, a data set's average is determined. Next the absolute deviation
between each data point and the average are determined. Thirdly, a rejection region is determined using the
formula:
;
where is the critical value from the Student t distribution with n-2 degrees of freedom, n is the sample
size, and s is the sample standard deviation. To determine if a value is an outlier: Calculate
. If δ > Rejection Region, the data point is an outlier. If δ ≤ Rejection Region, the data
point is not an outlier.
The modified Thompson Tau test is used to find one outlier at a time (largest value of δ is removed if it is an
outlier). Meaning, if a data point is found to be an outlier, it is removed from the data set and the test is
applied again with a new average and rejection region. This process is continued until no outliers remain in
a data set.
Some work has also examined outliers for nominal (or categorical) data. In the context of a set of examples
(or instances) in a data set, instance hardness measures the probability that an instance will be misclassified
( where y is the assigned class label and x represent the input attribute value for an instance in
the training set t).[25] Ideally, instance hardness would be calculated by summing over the set of all possible
hypotheses H:
Practically, this formulation is unfeasible as H is potentially infinite and calculating is unknown for
many algorithms. Thus, instance hardness can be approximated using a diverse subset :
where is the hypothesis induced by learning algorithm trained on training set t with
hyperparameters . Instance hardness provides a continuous value for determining if an instance is an
outlier instance.
Retention
Even when a normal distribution model is appropriate to the data being analyzed, outliers are expected for
large sample sizes and should not automatically be discarded if that is the case.[26] Instead, one should use a
method that is robust to outliers to model or analyze data with naturally occurring outliers.[26]
Exclusion
When deciding whether to remove an outlier, the cause has to be considered. As mentioned earlier, if the
outlier's origin can be attributed to an experimental error, or if it can be otherwise determined that the
outlying data point is erroneous, it is generally recommended to remove it.[26][27] However, it is more
desirable to correct the erroneous value, if possible.
Removing a data point solely because it is an outlier, on the other hand, is a controversial practice, often
frowned upon by many scientists and science instructors, as it typically invalidates statistical results.[26][27]
While mathematical criteria provide an objective and quantitative method for data rejection, they do not
make the practice more scientifically or methodologically sound, especially in small sets or where a normal
distribution cannot be assumed. Rejection of outliers is more acceptable in areas of practice where the
underlying model of the process being measured and the usual distribution of measurement error are
confidently known.
The two common approaches to exclude outliers are truncation (or trimming) and Winsorising. Trimming
discards the outliers whereas Winsorising replaces the outliers with the nearest "nonsuspect" data.[28]
Exclusion can also be a consequence of the measurement process, such as when an experiment is not
entirely capable of measuring such extreme values, resulting in censored data.[29]
In regression problems, an alternative approach may be to only exclude points which exhibit a large degree
of influence on the estimated coefficients, using a measure such as Cook's distance.[30]
If a data point (or points) is excluded from the data analysis, this should be clearly stated on any subsequent
report.
Non-normal distributions
The possibility should be considered that the underlying distribution of the data is not approximately
normal, having "fat tails". For instance, when sampling from a Cauchy distribution,[31] the sample variance
increases with the sample size, the sample mean fails to converge as the sample size increases, and outliers
are expected at far larger rates than for a normal distribution. Even a slight difference in the fatness of the
tails can make a large difference in the expected number of extreme values.
Set-membership uncertainties
A set membership approach considers that the uncertainty corresponding to the ith measurement of an
unknown random vector x is represented by a set Xi (instead of a probability density function). If no outliers
occur, x should belong to the intersection of all Xi's. When outliers occur, this intersection could be empty,
and we should relax a small number of the sets Xi (as small as possible) in order to avoid any
inconsistency.[32] This can be done using the notion of q-relaxed intersection. As illustrated by the figure,
the q-relaxed intersection corresponds to the set of all x which belong to all sets except q of them. Sets Xi
that do not intersect the q-relaxed intersection could be suspected to be outliers.
Alternative models
In cases where the cause of the outliers is known, it may be possible
to incorporate this effect into the model structure, for example by
using a hierarchical Bayes model, or a mixture model.[33][34]
See also
Anomaly (natural sciences)
Novelty detection
Anscombe's quartet
Data transformation (statistics) Figure 5. q-relaxed intersection of 6
Extreme value theory sets for q=2 (red), q=3 (green), q= 4
Influential observation (blue), q= 5 (yellow).
References
1. Grubbs, F. E. (February 1969). "Procedures for detecting outlying observations in samples".
Technometrics. 11 (1): 1–21. doi:10.1080/00401706.1969.10490657 (https://doi.org/10.108
0%2F00401706.1969.10490657). "An outlying observation, or "outlier," is one that appears
to deviate markedly from other members of the sample in which it occurs."
2. Maddala, G. S. (1992). "Outliers" (https://books.google.com/books?id=nBS3AAAAIAAJ&pg=
PA89). Introduction to Econometrics (https://archive.org/details/introductiontoec00madd/pag
e/89) (2nd ed.). New York: MacMillan. pp. 89 (https://archive.org/details/introductiontoec00m
add/page/89). ISBN 978-0-02-374545-4. "An outlier is an observation that is far removed
from the rest of the observations."
3. Pimentel, M. A., Clifton, D. A., Clifton, L., & Tarassenko, L. (2014). A review of novelty
detection. Signal Processing, 99, 215-249.
4. Grubbs 1969, p. 1 stating "An outlying observation may be merely an extreme manifestation
of the random variability inherent in the data. ... On the other hand, an outlying observation
may be the result of gross deviation from prescribed experimental procedure or an error in
calculating or recording the numerical value."
5. Ripley, Brian D. 2004. Robust statistics (http://www.stats.ox.ac.uk/pub/StatMeth/Robust.pdf)
Archived (https://web.archive.org/web/20121021081319/http://www.stats.ox.ac.uk/pub/StatM
eth/Robust.pdf) 2012-10-21 at the Wayback Machine
6. Chandan Mukherjee, Howard White, Marc Wuyts, 1998, "Econometrics and Data Analysis
for Developing Countries Vol. 1" [1] (https://books.google.com/books?id=H-lkYmatYtAC&dq
=median+is+less+precise+than+mean&pg=PA60)
7. Ruan, Da; Chen, Guoqing; Kerre, Etienne (2005). Wets, G. (ed.). Intelligent Data Mining:
Techniques and Applications (https://archive.org/details/intelligentdatam00ruan_742).
Studies in Computational Intelligence Vol. 5. Springer. p. 318 (https://archive.org/details/intel
ligentdatam00ruan_742/page/n326). ISBN 978-3-540-26256-5.
8. Zimek, Arthur; Filzmoser, Peter (2018). "There and back again: Outlier detection between
statistical reasoning and data mining algorithms" (https://findresearcher.sdu.dk:8443/ws/files/
153197807/There_and_Back_Again.pdf) (PDF). Wiley Interdisciplinary Reviews: Data
Mining and Knowledge Discovery. 8 (6): e1280. doi:10.1002/widm.1280 (https://doi.org/10.1
002%2Fwidm.1280). ISSN 1942-4787 (https://www.worldcat.org/issn/1942-4787).
S2CID 53305944 (https://api.semanticscholar.org/CorpusID:53305944).
9. Pimentel, M. A., Clifton, D. A., Clifton, L., & Tarassenko, L. (2014). A review of novelty
detection. Signal Processing, 99, 215-249.
10. Rousseeuw, P; Leroy, A. (1996), Robust Regression and Outlier Detection (3rd ed.), John
Wiley & Sons
11. Hodge, Victoria J.; Austin, Jim (2004), "A Survey of Outlier Detection Methodologies",
Artificial Intelligence Review, 22 (2): 85–126, CiteSeerX 10.1.1.109.1943 (https://citeseerx.is
t.psu.edu/viewdoc/summary?doi=10.1.1.109.1943),
doi:10.1023/B:AIRE.0000045502.10941.a9 (https://doi.org/10.1023%2FB%3AAIRE.000004
5502.10941.a9), S2CID 3330313 (https://api.semanticscholar.org/CorpusID:3330313)
12. Barnett, Vic; Lewis, Toby (1994) [1978], Outliers in Statistical Data (3 ed.), Wiley, ISBN 978-
0-471-93094-5
13. Zimek, A.; Schubert, E.; Kriegel, H.-P. (2012). "A survey on unsupervised outlier detection in
high-dimensional numerical data". Statistical Analysis and Data Mining. 5 (5): 363–387.
doi:10.1002/sam.11161 (https://doi.org/10.1002%2Fsam.11161). S2CID 6724536 (https://ap
i.semanticscholar.org/CorpusID:6724536).
14. E178: Standard Practice for Dealing With Outlying Observations (https://www.nrc.gov/docs/
ML1023/ML102371244.pdf)
15. Benjamin Peirce, "Criterion for the Rejection of Doubtful Observations" (http://articles.adsab
s.harvard.edu/cgi-bin/nph-iarticle_query?1852AJ......2..161P;data_type=PDF_HIGH),
Astronomical Journal II 45 (1852) and Errata to the original paper (http://articles.adsabs.harv
ard.edu/cgi-bin/nph-iarticle_query?1852AJ......2..176P;data_type=PDF_HIGH).
16. Peirce, Benjamin (May 1877 – May 1878). "On Peirce's criterion". Proceedings of the
American Academy of Arts and Sciences. 13: 348–351. doi:10.2307/25138498 (https://doi.or
g/10.2307%2F25138498). JSTOR 25138498 (https://www.jstor.org/stable/25138498).
17. Peirce, Charles Sanders (1873) [1870]. "Appendix No. 21. On the Theory of Errors of
Observation". Report of the Superintendent of the United States Coast Survey Showing the
Progress of the Survey During the Year 1870: 200–224.. NOAA PDF Eprint (http://docs.lib.no
aa.gov/rescue/cgs/001_pdf/CSC-0019.PDF#page=215) (goes to Report p. 200, PDF's p.
215).
18. Peirce, Charles Sanders (1986) [1982]. "On the Theory of Errors of Observation". In Kloesel,
Christian J. W.; et al. (eds.). Writings of Charles S. Peirce: A Chronological Edition (https://ar
chive.org/details/writingsofcharle0002peir/page/140). Vol. 3, 1872–1878. Bloomington,
Indiana: Indiana University Press. pp. 140–160 (https://archive.org/details/writingsofcharle00
02peir/page/140). ISBN 978-0-253-37201-7. – Appendix 21, according to the editorial note
on page 515
19. Tukey, John W (1977). Exploratory Data Analysis (https://archive.org/details/exploratorydata
a00tuke_0). Addison-Wesley. ISBN 978-0-201-07616-5. OCLC 3058187 (https://www.worldc
at.org/oclc/3058187).
20. Knorr, E. M.; Ng, R. T.; Tucakov, V. (2000). "Distance-based outliers: Algorithms and
applications". The VLDB Journal the International Journal on Very Large Data Bases. 8 (3–
4): 237. CiteSeerX 10.1.1.43.1842 (https://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.
1.43.1842). doi:10.1007/s007780050006 (https://doi.org/10.1007%2Fs007780050006).
S2CID 11707259 (https://api.semanticscholar.org/CorpusID:11707259).
21. Ramaswamy, S.; Rastogi, R.; Shim, K. (2000). Efficient algorithms for mining outliers from
large data sets. Proceedings of the 2000 ACM SIGMOD international conference on
Management of data - SIGMOD '00. p. 427. doi:10.1145/342009.335437 (https://doi.org/10.1
145%2F342009.335437). ISBN 1581132174.
22. Breunig, M. M.; Kriegel, H.-P.; Ng, R. T.; Sander, J. (2000). LOF: Identifying Density-based
Local Outliers (http://www.dbs.ifi.lmu.de/Publikationen/Papers/LOF.pdf) (PDF). Proceedings
of the 2000 ACM SIGMOD International Conference on Management of Data. SIGMOD.
pp. 93–104. doi:10.1145/335191.335388 (https://doi.org/10.1145%2F335191.335388).
ISBN 1-58113-217-4.
23. Schubert, E.; Zimek, A.; Kriegel, H. -P. (2012). "Local outlier detection reconsidered: A
generalized view on locality with applications to spatial, video, and network outlier
detection". Data Mining and Knowledge Discovery. 28: 190–237. doi:10.1007/s10618-012-
0300-z (https://doi.org/10.1007%2Fs10618-012-0300-z). S2CID 19036098 (https://api.sema
nticscholar.org/CorpusID:19036098).
24. Thompson .R. (1985). "A Note on Restricted Maximum Likelihood Estimation with an
Alternative Outlier Model (https://www.jstor.org/stable/2345543?seq=1#page_scan_tab_cont
ents)".Journal of the Royal Statistical Society. Series B (Methodological), Vol. 47, No. 1, pp.
53-55
25. Smith, M.R.; Martinez, T.; Giraud-Carrier, C. (2014). "An Instance Level Analysis of Data
Complexity (https://link.springer.com/article/10.1007%2Fs10994-013-5422-z)". Machine
Learning, 95(2): 225-256.
26. Karch, Julian D. (2023). "Outliers may not be automatically removed". "Journal of
Experimental Psychology: General". doi:10.1037/xge0001357 (https://doi.org/10.1037%2Fx
ge0001357).
27. Bakker, Marjan; Wicherts, Jelte M. (2014). "Outlier removal, sum scores, and the inflation of
the type I error rate in independent samples t tests: The power of alternatives and
recommendations". Psychological Methods. 19 (3): 409–427. doi:10.1037/met0000014 (http
s://doi.org/10.1037%2Fmet0000014).
28. Wike, Edward L. (2006). Data Analysis: A Statistical Primer for Psychology Students. pp. 24–
25. ISBN 9780202365350.
29. Dixon, W. J. (June 1960). "Simplified estimation from censored normal samples" (http://proje
cteuclid.org/download/pdf_1/euclid.aoms/1177705900). The Annals of Mathematical
Statistics. 31 (2): 385–391. doi:10.1214/aoms/1177705900 (https://doi.org/10.1214%2Faom
s%2F1177705900).
30. Cook, R. Dennis (Feb 1977). "Detection of Influential Observations in Linear Regression".
Technometrics (American Statistical Association) 19 (1): 15–18.
31. Weisstein, Eric W. Cauchy Distribution. From MathWorld--A Wolfram Web Resource (http://m
athworld.wolfram.com/CauchyDistribution.html)
32. Jaulin, L. (2010). "Probabilistic set-membership approach for robust regression" (http://www.
ensta-bretagne.fr/jaulin/paper_probint_0.pdf) (PDF). Journal of Statistical Theory and
Practice. 4: 155–167. doi:10.1080/15598608.2010.10411978 (https://doi.org/10.1080%2F15
598608.2010.10411978). S2CID 16500768 (https://api.semanticscholar.org/CorpusID:16500
768).
33. Roberts, S. and Tarassenko, L.: 1995, A probabilistic resource allocating network for novelty
detection. Neural Computation 6, 270–284.
34. Bishop, C. M. (August 1994). "Novelty detection and Neural Network validation". IEE
Proceedings - Vision, Image, and Signal Processing. 141 (4): 217–222. doi:10.1049/ip-
vis:19941330 (https://doi.org/10.1049%2Fip-vis%3A19941330).
External links
Renze, John. "Outlier" (https://mathworld.wolfram.com/Outlier.html). MathWorld.
Balakrishnan, N.; Childs, A. (2001) [1994], "Outlier" (https://www.encyclopediaofmath.org/ind
ex.php?title=Outlier), Encyclopedia of Mathematics, EMS Press
Grubbs test (http://www.itl.nist.gov/div898/handbook/eda/section3/eda35h.htm) described by
NIST manual