This article gives a scintillation detector fault diagnosis method based on BP neural network. From the aspect of output signals of scintillation detectors, the wavelet packet transform is used to extract the energy characteristic vectors which are treated as the input of BP neural network, and a training database is established as well as BP neural network parameters are optimized. Then the method is employed to establish a fault recognition model and fault types can be concluded. Finally, the simulation data are compared with those of two other methods (the statistical diagnosis method and an method based on multi-classification support vector machine). The experimental results illustrate that the application of proposed method can improve the fault diagnosis accuracy of scintillation detectors effectively.

This paper presents a scintillation detector fault diagnosis method based on wavelet packet analysis and multi-classification support vector machine(multi-SVM). In the proposed method, a wavelet packet algorithm is used to analyze waveform characteristics of output signals caused by different faults of detectors, then characteristic vectors can be extracted. The multi-SVM is employed to establish a fault recognition model and fault types can be concluded. Performances of the proposed method are validated by experimental data obtained from the plastic scintillation detector for the single fault diagnosis and hybrid fault diagnosis. The experimental results show that faults can be diagnosed automatically and quickly by analyzing signal waveform features. K: Analysis and statistical methods; Models and simulations; Radiation monitoring; Gamma detectors (scintillators, CZT, HPGe, HgI etc) 1Corresponding author.

No abstract

Outlier detection is critical in real world. Due to the existence of many outlier detection techniques which often return different results for the same data set, the users have to address the problem of determining which among these techniques is the best suited for their task and tune its parameters. This is particularly challenging in the unsupervised setting, where no labels are available for cross-validation needed for such method and parameter optimization. In this work, we propose AutoOD which uses the existing unsupervised detection techniques to automatically produce high quality outliers without any human tuning. AutoOD's fundamentally new strategy unifies the merits of unsupervised outlier detection and supervised classification within one integrated solution. It automatically tests a diverse set of unsupervised outlier detectors on a target data set, extracts useful signals from their combined detection results to reliably capture key differences between outliers and inliers. It then uses these signals to produce a "custom outlier classifier" to classify outliers, with its accuracy comparable to supervised outlier classification models trained with ground truth labels - without having access to the much needed labels. On a diverse set of benchmark outlier detection datasets, AutoOD consistently outperforms the best unsupervised outlier detector selected from hundreds of detectors. It also outperforms other tuning-free approaches from 12 to 97 points (out of 100) in the F-1 score.

No abstract