Score (statistics)

In statistics, the score (or informant[1]) is the gradient of the log-likelihood function with respect to the
parameter vector. Evaluated at a particular point of the parameter vector, the score indicates the steepness of
the log-likelihood function and thereby the sensitivity to infinitesimal changes to the parameter values. If the
log-likelihood function is continuous over the parameter space, the score will vanish at a local maximum or
minimum; this fact is used in maximum likelihood estimation to find the parameter values that maximize the
likelihood function.

Since the score is a function of the observations that are subject to sampling error, it lends itself to a test
statistic known as score test in which the parameter is held at a particular value. Further, the ratio of two
likelihood functions evaluated at two distinct parameter values can be understood as a definite integral of
the score function.[2]

Definition
The score is the gradient (the vector of partial derivatives) of , the natural logarithm of the
likelihood function, with respect to an m-dimensional parameter vector .

This differentiation yields a row vector, and indicates the sensitivity of the likelihood (its
derivative normalized by its value).

In older literature, "linear score" may refer to the score with respect to infinitesimal translation of a given
density. This convention arises from a time when the primary parameter of interest was the mean or median
of a distribution. In this case, the likelihood of an observation is given by a density of the form
. The "linear score" is then defined as

Properties

Mean

While the score is a function of , it also depends on the observations at which the
likelihood function is evaluated, and in view of the random character of sampling one may take its expected
value over the sample space. Under certain regularity conditions on the density functions of the random
variables,[3][4] the expected value of the score, evaluated at the true parameter value , is zero. To see this,
rewrite the likelihood function as a probability density function , and denote the
sample space . Then:
The assumed regularity conditions allow the interchange of derivative and integral (see Leibniz integral
rule), hence the above expression may be rewritten as

It is worth restating the above result in words: the expected value of the score, at true parameter value is
zero. Thus, if one were to repeatedly sample from some distribution, and repeatedly calculate the score,
then the mean value of the scores would tend to zero asymptotically.

Variance

The variance of the score, , can be derived from the above expression for the
expected value.

Hence the variance of the score is equal to the negative expected value of the Hessian matrix of the log-
likelihood.[5]
The latter is known as the Fisher information and is written . Note that the Fisher information is not a
function of any particular observation, as the random variable has been averaged out. This concept of
information is useful when comparing two methods of observation of some random process.

Examples

Bernoulli process

Consider observing the first n trials of a Bernoulli process, and seeing that A of them are successes and the
remaining B are failures, where the probability of success is θ.

Then the likelihood is

so the score s is

We can now verify that the expectation of the score is zero. Noting that the expectation of A is nθ and the
expectation of B is n(1 − θ) [recall that A and B are random variables], we can see that the expectation of s
is

We can also check the variance of . We know that A + B = n (so B = n − A) and the variance of A is
nθ(1 − θ) so the variance of s is

Binary outcome model

For models with binary outcomes (Y = 1 or 0), the model can be scored with the logarithm of predictions

where p is the probability in the model to be estimated and S is the score.[6]

Applications

Scoring algorithm
The scoring algorithm is an iterative method for numerically determining the maximum likelihood estimator.

Score test

Note that is a function of and the observation , so that, in general, it is not a

Further note that the likelihood-ratio test is given by

which means that the likelihood-ratio test can be understood as the area under the score function between
and .[8]

Score matching (machine learning)

Score matching describes the process of applying machine learning algorithms (commonly neural
networks) to approximate the score function of an unknown distribution from
finite samples. The learned function can then be used in generative modeling to draw new samples from
.[9]

It might seem confusing that the word score has been used for , because it is not a likelihood
function, neither it has a derivative with respect to the parameters. For more information about this
definition, see the referenced paper. [10]

See also
Fisher information – Notion in statistics
Information theory – Scientific study of digital information
Score test – Statistical test based on the gradient of the likelihood function
Scoring algorithm – form of Newton's method used in statistics
Standard score – How many standard deviations apart from the mean an observed datum is
Support curve – Function related to statistics and probability theory

Notes
1. Informant in Encyclopaedia of Maths (https://encyclopediaofmath.org/wiki/Informant)
2. Pickles, Andrew (1985). An Introduction to Likelihood Analysis (https://archive.org/details/intr
oductiontoli0000pick/page/24). Norwich: W. H. Hutchins & Sons. pp. 24–29 (https://archive.o
rg/details/introductiontoli0000pick/page/24). ISBN 0-86094-190-6.
3. Serfling, Robert J. (1980). Approximation Theorems of Mathematical Statistics (https://archiv
e.org/details/approximationthe00serf). New York: John Wiley & Sons. p. 145 (https://archive.
 org/details/approximationthe00serf/page/n162). ISBN 0-471-02403-1.
4. Greenberg, Edward; Webster, Charles E. Jr. (1983). Advanced Econometrics : A Bridge to
the Literature (https://books.google.com/books?id=TSK7AAAAIAAJ&pg=PA25). New York:
John Wiley & Sons. p. 25. ISBN 0-471-09077-8.
5. Sargan, Denis (1988). Lectures on Advanced Econometrics. Oxford: Basil Blackwell.
pp. 16–18. ISBN 0-631-14956-2.
6. Steyerberg, E. W.; Vickers, A. J.; Cook, N. R.; Gerds, T.; Gonen, M.; Obuchowski, N.;
Pencina, M. J.; Kattan, M. W. (2010). "Assessing the performance of prediction models. A
framework for traditional and novel measures" (https://www.ncbi.nlm.nih.gov/pmc/articles/PM
C3575184). Epidemiology. 21 (1): 128–138. doi:10.1097/EDE.0b013e3181c30fb2 (https://do
i.org/10.1097%2FEDE.0b013e3181c30fb2). PMC 3575184 (https://www.ncbi.nlm.nih.gov/p
mc/articles/PMC3575184). PMID 20010215 (https://pubmed.ncbi.nlm.nih.gov/20010215).
7. Rao, C. Radhakrishna (1948). "Large sample tests of statistical hypotheses concerning
several parameters with applications to problems of estimation". Mathematical Proceedings
of the Cambridge Philosophical Society. 44 (1): 50–57. Bibcode:1948PCPS...44...50R (http
s://ui.adsabs.harvard.edu/abs/1948PCPS...44...50R). doi:10.1017/S0305004100023987 (htt
ps://doi.org/10.1017%2FS0305004100023987). S2CID 122382660 (https://api.semanticsch
olar.org/CorpusID:122382660).
8. Buse, A. (1982). "The Likelihood Ratio, Wald, and Lagrange Multiplier Tests: An Expository
Note". The American Statistician. 36 (3a): 153–157. doi:10.1080/00031305.1982.10482817
(https://doi.org/10.1080%2F00031305.1982.10482817).
9. Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon,
Ben Poole (2020). "Score-Based Generative Modeling through Stochastic Differential
Equations". arXiv:2011.13456 (https://arxiv.org/abs/2011.13456) [cs.LG (https://arxiv.org/arch
ive/cs.LG)].
10. https://www.jmlr.org/papers/volume6/hyvarinen05a/hyvarinen05a.pdf

References
Chentsov, N.N. (2001) [1994], "Informant" (https://www.encyclopediaofmath.org/index.php?tit
le=Informant), Encyclopedia of Mathematics, EMS Press
Cox, D. R.; Hinkley, D. V. (1974). Theoretical Statistics. Chapman & Hall. ISBN 0-412-
12420-3.
Schervish, Mark J. (1995). Theory of Statistics. New York: Springer. Section 2.3.1. ISBN 0-
387-94546-6.

Retrieved from "https://en.wikipedia.org/w/index.php?title=Score_(statistics)&oldid=1165079345"