Frequency Band Selection Based on a New Indicator: Accuracy Rate

2017 ◽  
Author(s):  
Yaqiang Jin ◽  
Zhiliang Liu ◽  
Jinlong Kang ◽  
Jie Yin ◽  
Dandan Peng
Keyword(s):  
Frequency Band ◽  
Cross Validation ◽  
Nearest Neighbor ◽  
Diagnostic Methods ◽  
Center Frequency ◽  
Accuracy Rate ◽  
Nearest Neighbor Classifier ◽  
Envelope Spectrum ◽  
Local Defects ◽  
Neighbor Classifier

Local defects in rotating machinery give rise to periodic impulses in vibrations. In order to acquire the information of these faults, various diagnostic methods have been proposed in the past decades. Most methods used the squared envelope spectrum (i.e., the spectrum of the squared envelope) as the final diagnostic tool, but different preprocessing steps were used before obtaining the envelope signal. The key problem is to obtain the center frequency and bandwidth of the fault signal, then analyze the envelope (squared envelope spectrum) of the band-pass filtered signal. The framework of accuracy rate method was proposed by means of cross validation of the nearest neighbor classifier in this paper: a) obtain the piecewise signal through original signal segmentation; b) calculate the feature of each piecewise signal; c) then an accuracy rate is calculated based on cross validation of the nearest neighbor classifier; and d) repeat the above steps in different frequency band, then find a frequency band with the maximum accuracy rate. Through this algorithm, we can obtain a fault frequency band, and then we can find out the type of the fault by the spectrum of the squared envelope. At the end of this paper, the proposed method is validated by two examples and compared with the other two diagnostic methods: conventional envelope analysis and Fast kurtogram. Through the comparison of results, the validity and superiority of this method has been proved.

scholarly journals Dressing Tool Condition Monitoring through Impedance-Based Sensors: Part 2—Neural Networks and K-Nearest Neighbor Classifier Approach

Sensors ◽  
2018 ◽  
Vol 18 (12) ◽  
pp. 4453 ◽  
Author(s):  
Pedro Junior ◽  
Doriana D’Addona ◽  
Paulo Aguiar ◽  
Roberto Teti
Keyword(s):  
Neural Networks ◽  
Frequency Band ◽  
Nearest Neighbor ◽  
K Nearest Neighbor ◽  
Electromechanical Impedance ◽  
Nearest Neighbor Classifier ◽  
Tool Condition ◽  
Dressing Tool ◽  
Sensor Monitoring ◽  
Neighbor Classifier

This paper presents an approach for impedance-based sensor monitoring of dressing tool condition in grinding by using the electromechanical impedance (EMI) technique. This method was introduced in Part 1 of this work and the purpose of this paper (Part 2) is to achieve an optimal selection of the excitation frequency band based on multi-layer neural networks (MLNN) and k-nearest neighbor classifier (k-NN). The proposed approach was validated on the basis of dressing tool condition information obtained from the monitoring of experimental dressing tests with two industrial stationary single-point dressing tools. Moreover, representative damage indices for diverse damage cases, obtained from impedance signatures at different frequency bands, were taken into account for MLNN data processing. The intelligent system was able to select the most damage-sensitive features based on optimal frequency band. The best models showed a general overall error lower than 2%, thus robustly contributing to the efficient automation of grinding and dressing operations. The promising results of this study foster the EMI-based sensor monitoring approach to fault diagnosis in dressing operations and its effective implementation for industrial grinding process automation.

Diagnostics of Synchronous Motor Based on Analysis of Acoustic Signals with the use of Line Spectral Frequencies and K-nearest Neighbor Classifier

2015 ◽  
Vol 39 (2) ◽  
pp. 189-194 ◽  
Author(s):  
Adam Glowacz
Keyword(s):  
Nearest Neighbor ◽  
Acoustic Signals ◽  
Synchronous Motor ◽  
Diagnostic Methods ◽  
Mining Equipment ◽  
K Nearest Neighbor ◽  
Nearest Neighbor Classifier ◽  
Fault Diagnostic ◽  
Electrical Motors ◽  
Neighbor Classifier

Abstract In industrial processes electrical motors are serviced after a specific number of hours, even if there is a need for service. This led to the development of early fault diagnostic methods. Paper presents early fault diagnostic method of synchronous motor. This method uses acoustic signals generated by synchronous motor. Plan of study of acoustic signal of synchronous motor was proposed. Two conditions of synchronous motor were analyzed. Studies were carried out for methods of data processing: Line Spectral Frequencies and K-Nearest Neighbor classifier with Minkowski distance. Condition monitoring is useful to protect electric motors and mining equipment. In the future, these studies can be used in other electrical devices.

scholarly journals Stronger Automation for Flyspeck by Feature Weighting and Strategy Evolution

10.29007/5gzr ◽  
2018 ◽  
Author(s):  
Cezary Kaliszyk ◽  
Josef Urban
Keyword(s):  
Nearest Neighbor ◽  
Feature Weighting ◽  
K Nearest Neighbor ◽  
Nearest Neighbor Classifier ◽  
Hol Light ◽  
Distance Weighted ◽  
Neighbor Classifier

Two complementary AI methods are used to improve the strength of the AI/ATP service for proving conjectures over the HOL Light and Flyspeck corpora. First, several schemes for frequency-based feature weighting are explored in combination with distance-weighted k-nearest-neighbor classifier. This results in 16% improvement (39.0% to 45.5% Flyspeck problems solved) of the overall strength of the service when using 14 CPUs and 30 seconds. The best premise-selection/ATP combination is improved from 24.2% to 31.4%, i.e. by 30%. A smaller improvement is obtained by evolving targetted E prover strategies on two particular premise selections, using the Blind Strategymaker (BliStr) system. This raises the performance of the best AI/ATP method from 31.4% to 34.9%, i.e. by 11%, and raises the current 14-CPU power of the service to 46.9%.

scholarly journals A patient distance metric for neurology

2020 ◽  
Author(s):  
Daniel B Hier ◽  
Jonathan Kopel ◽  
Steven U Brint ◽  
Donald C Wunsch II ◽  
Gayla R Olbricht ◽  
...  
Keyword(s):  
Nearest Neighbor ◽  
Signs And Symptoms ◽  
Diagnostic Error ◽  
K Nearest Neighbor ◽  
Bipartite Matching ◽  
Nearest Neighbor Classifier ◽  
Neurological Signs ◽  
Neurological Patients ◽  
Machine Readable ◽  
Neighbor Classifier

Abstract Objective: Neurologists lack a metric for measuring the distance between neurological patients. When neurological signs and symptoms are represented as neurological concepts from a hierarchical ontology and neurological patients are represented as sets of concepts, distances between patients can be represented as inter-set distances.Methods:We converted the neurological signs and symptoms from 721 published neurology cases into sets of concepts with corresponding machine-readable codes. We calculated inter-concept distances based a hierarchical ontology and we calculated inter-patient distances by semantic weighted bipartite matching. We evaluated the accuracy of a k-nearest neighbor classifier to allocate patients into 40 diagnostic classes.Results:Within a given diagnosis, mean patient distance differed by diagnosis, suggesting that across diagnoses there are differences in how similar patients are to other patients with the same diagnosis. The mean distance from one diagnosis to another diagnosis differed by diagnosis, suggesting that diagnoses differ in their proximity to other diagnoses. Utilizing a k-nearest neighbor classifier and inter-patient distances, the risk of misclassification differed by diagnosis.Conclusion:If signs and symptoms are converted to machine-readable codes and patients are represented as sets of these codes, patient distances can be computed as an inter-set distance. These patient distances given insights into how homogeneous patients are within a diagnosis (stereotypy), the distance between different diagnoses (proximity), and the risk of diagnosis misclassification (diagnostic error).

Sign in / Sign up

Export Citation Format

Share Document