Consider the unit square in the Euclidean plane. Consider the probability measure defined on by the restriction of two-dimensional Lebesgue measure to . That is, the probability of an event is simply the area of . We assume is a measurable subset of .
Consider a one-dimensional subset of such as the line segment . has -measure zero; every subset of is a -null set; since the Lebesgue measure space is a complete measure space,
While true, this is somewhat unsatisfying. It would be nice to say that "restricted to" is the one-dimensional Lebesgue measure , rather than the zero measure. The probability of a "two-dimensional" event could then be obtained as an integral of the one-dimensional probabilities of the vertical "slices" : more formally, if denotes one-dimensional Lebesgue measure on , then
for any "nice" . The disintegration theorem makes this argument rigorous in the context of measures on metric spaces.
Statement of the theoremedit
(Hereafter, will denote the collection of Borel probability measures on a topological space.)
The assumptions of the theorem are as follows:
Let be a Borel-measurable function. Here one should think of as a function to "disintegrate" , in the sense of partitioning into . For example, for the motivating example above, one can define , , which gives that , a slice we want to capture.
Let be the pushforward measure. This measure provides the distribution of (which corresponds to the events ).
The conclusion of the theorem: There exists a -almost everywhere uniquely determined family of probability measures , which provides a "disintegration" of into , such that:
the function is Borel measurable, in the sense that is a Borel-measurable function for each Borel-measurable set ;
The original example was a special case of the problem of product spaces, to which the disintegration theorem applies.
When is written as a Cartesian product and is the natural projection, then each fibre can be canonically identified with and there exists a Borel family of probability measures in (which is -almost everywhere uniquely determined) such that
The disintegration theorem can also be seen as justifying the use of a "restricted" measure in vector calculus. For instance, in Stokes' theorem as applied to a vector field flowing through a compactsurface, it is implicit that the "correct" measure on is the disintegration of three-dimensional Lebesgue measure on , and that the disintegration of this measure on ∂Σ is the same as the disintegration of on .[2]
Conditional distributionsedit
The disintegration theorem can be applied to give a rigorous treatment of conditional probability distributions in statistics, while avoiding purely abstract formulations of conditional probability.[3]
^Dellacherie, C.; Meyer, P.-A. (1978). Probabilities and Potential. North-Holland Mathematics Studies. Amsterdam: North-Holland. ISBN 0-7204-0701-X.
^Ambrosio, L.; Gigli, N.; Savaré, G. (2005). Gradient Flows in Metric Spaces and in the Space of Probability Measures. ETH Zürich, Birkhäuser Verlag, Basel. ISBN 978-3-7643-2428-5.
^Chang, J.T.; Pollard, D. (1997). "Conditioning as disintegration" (PDF). Statistica Neerlandica. 51 (3): 287. CiteSeerX10.1.1.55.7544. doi:10.1111/1467-9574.00056. S2CID 16749932.