Asset-specific bayesian diagnostics in mixed contexts

Abstract

Abstract-In this paper we build upon previous work to examine the efficacy of blending probabilities in asset-specific classifiers to improve diagnostic accuracy for a fleet of assets. In previous work we also introduced the idea of using split probabilities. We add environmental differentiation to asset differentiation in the experiments and assume that data is acquired in the context of online health monitoring. We hypothesize that overall diagnostic accuracy will be increased with the blending approach relative to the single aggregate classifier or split probability assetspecific classifiers. The hypothesis is largely supported by the results. Future work will concentrate on improving the blending mechanism and working with small data sets.

Distributional Smoothing in Bayesian Fault Diagnosis

“…One of the areas we left for future research in our previous paper [8] was a sensitivity analysis of our results for different values of the user defined parameters k and q in our implementation of Equation 3. In the results just presented, we used of k = 100 and q = 1.2.…”

“…The resulting function will cause m to decrease as noise and amount of data increases. This formula represents a generalization of the result provided in [8]. Consequently, several options now exist for f(o(T i ),D j ).…”

“…Examples of the different DMatrices-reflecting different usage profiles or environmental conditions-included the baseline D-Matrix and D-Matrices where one test always failed and another where one test always passed. All variations are described fully in [8].…”

“…Choosing a diagnosis at random broke all classification ties. For a more in-depth explanation of the experimental design, see our previous papers [1], [2], [8].…”