Truncated Gaussians as Tolerance Sets
نویسندگان
چکیده
This work focuses on the use of truncated Gaussian distributions as models for bounded data { measurements that are constrained to appear between xed limits. We prove that the truncated Gaussian can be viewed as a maximum entropy distribution for truncated bounded data, when mean and covariance are given. We present the characteristic function for the truncated Gaussian; from this, we derive algorithms for calculation of mean, variance, summation, application of Bayes rule and ltering with truncated Gaussians. As an example of the power of our methods, we describe a derivation of the disparity constraint (used in computer vision) from our models. Our approach complements results in Statistics, but our proposal is not only to use the truncated Gaussian as a model for selected data; we propose to model measurements as fundamentally bounded in terms of truncated Gaussians.
منابع مشابه
Maximum likelihood method for parameter estimation of bell-shaped functions on graphs
To estimate the location-scale parameters of a bell-shaped density on attributed graphs, we consider radial densities as approximations. The problem of estimating the parameters of radial densities on graphs is equivalent to the problem of estimating the parameters of truncated Gaussians in a Euclidean space. Based on this result, we adopt the maximum likelihood method for truncated Gaussians. ...
متن کاملBandlimited Approximations to the Truncated Gaussian and Applications
In this paper we extend the theory of optimal approximations of functions f : R → R in the L1(R)-metric by entire functions of prescribed exponential type (bandlimited functions). We solve this problem for the truncated and the odd Gaussians using explicit integral representations and properties of truncated theta functions obtained via the maximum principle for the heat operator. As applicatio...
متن کاملEnhancing the Interactive Visualization of Procedurally Encoded Multifield Data with Ellipsoidal Basis Functions
Functional approximation of scattered data is a popular technique for compactly representing various types of datasets in computer graphics, including surface, volume, and vector datasets. Typically, sums of Gaussians or similar radial basis functions are used in the functional approximation and PC graphics hardware is used to quickly evaluate and render these datasets. Previously, researchers ...
متن کاملDesign of compact acoustic models through clustering of tied-covariance Gaussians
We propose a new approach for designing compact acoustic models particularly suited to large systems that combine multiple model sets to represent distinct acoustic conditions or languages. We show that Gaussians based on mixtures of inverse covariances (MIC) with shared parameters can be clustered using an efficient Lloyd algorithm. As a result, more compact acoustic models can be built by clu...
متن کاملPros and Cons of truncated Gaussian EP in the context of Approximate Inference Control
Approximate Inference Control (AICO, [1]) is a method for solving Stochastic Optimal Control (SOC) problems. A core step in this method is to approximate the belief in each time step, that is, compute a distribution over the system state given the coupling to the previous state, the next state, and the current constraints or goals. Since constraints and goals may be arbitrarily non-linear (spec...
متن کامل