MATLAB tool for probability density assessment and nonparametric estimation
Jenny Farmer, Donald J. Jacobs
Abstract
A MATLAB function is presented for nonparametric probability density estimation, based on an iterative method that employs the principle of maximum entropy and characteristic properties of single order statistics. Featuring a robust and adaptive design, the method is well-suited for high throughput applications. The implementation comprises a MATLAB interface and underlying C++ code with extensible components that can be easily integrated into third party software. The functionality includes plotting capabilities and model independent diagnostics featuring the scaled quantile residual that is invariant to sample size, distribution, and estimation method.
Topics & Concepts
MATLABComputer scienceNonparametric statisticsQuantileDensity estimationProbability density functionSoftwareProbability distributionResidualPrinciple of maximum entropyAlgorithmSource codeQuantile functionInterface (matter)Entropy (arrow of time)Data miningCumulative distribution functionStatisticsArtificial intelligenceMathematicsProgramming languageParallel computingQuantum mechanicsBubbleEstimatorPhysicsMaximum bubble pressure methodGaussian Processes and Bayesian InferenceAdvanced Statistical Methods and ModelsControl Systems and Identification