Methods to integrate multinormals and compute classification measures (2012.14331v12)
Abstract: Univariate and multivariate normal probability distributions are widely used when modeling decisions under uncertainty. Computing the performance of such models requires integrating these distributions over specific domains, which can vary widely across models. Besides some special cases, there exist no general analytical expressions, standard numerical methods or software for these integrals. Here we present mathematical results and open-source software that provide (i) the probability in any domain of a normal in any dimensions with any parameters, (ii) the probability density, cumulative distribution, and inverse cumulative distribution of any function of a normal vector, (iii) the classification errors among any number of normal distributions, the Bayes-optimal discriminability index and relation to the operating characteristic, (iv) ways to scale the discriminability of two distributions, (v) dimension reduction and visualizations for such problems, and (vi) tests for how reliably these methods may be used on given data. We demonstrate these tools with vision research applications of detecting occluding objects in natural scenes, and detecting camouflage.
- A method to integrate and classify normal distributions. Journal of Vision, 21(10):1–1, 2021.
- Andrew Ng. Generative learning algorithms. CS229 Lecture notes, IV, 2019.
- Signal detection theory and psychophysics, volume 1. Wiley New York, 1966.
- Pattern classification. John Wiley & Sons, 2012.
- Harold Ruben. Probability content of regions under spherical normal distributions, i. The Annals of Mathematical Statistics, 31(3):598–618, 1960.
- Comparison of methods for the computation of multivariate t probabilities. Journal of Computational and Graphical Statistics, 11(4):950–971, 2002.
- Computation of multivariate normal and t probabilities, volume 195. Springer Science & Business Media, 2009.
- Harold Ruben. Probability content of regions under spherical normal distributions, iv: The distribution of homogeneous and non-homogeneous quadratic functions of normal variables. The Annals of Mathematical Statistics, 33(2):542–570, 1962.
- Robert B Davies. Numerical inversion of a characteristic function. Biometrika, 60(2):415–417, 1973.
- Lloyd N Trefethen. Approximation theory and approximation practice, volume 164. Siam, 2019.
- Distributing many points on a sphere. The mathematical intelligencer, 19(1):5–11, 1997.
- What is the best index of detectability? Psychological Bulletin, 80(6):481, 1973.
- David M Green. A homily on signal detection theory. The Journal of the Acoustical Society of America, 148(1):222–225, 2020.
- Thomas D Wickens. Elementary signal detection theory. Oxford University Press, USA, 2002.
- RL Chaddha and LF Marcus. An empirical comparison of distance statistics for populations with unequal covariance matrices. Biometrics, pages 683–694, 1968.
- SA Paranjpe and AP Gore. Selecting variables for discrimination when covariance matrices are unequal. Statistics & probability letters, 21(5):417–419, 1994.
- Psychophysics and signal detection. Technical report, Indiana University Hearing and Communication Laboratory, 1962.
- Weighted linear cue combination with possibly correlated error. Vision research, 43(23):2451–2468, 2003.
- Detection of occluding targets in natural backgrounds. Journal of Vision, 20(13):14, 2020.
- Understanding camouflage detection. Journal of Vision, 18(10):549, 2018.