The pruning algorithms for sparse least squares support vector regression machine are common methods, and easily com- prehensible, but the computational burden in the training phase is heavy due to the retraining in p...The pruning algorithms for sparse least squares support vector regression machine are common methods, and easily com- prehensible, but the computational burden in the training phase is heavy due to the retraining in performing the pruning process, which is not favorable for their applications. To this end, an im- proved scheme is proposed to accelerate sparse least squares support vector regression machine. A major advantage of this new scheme is based on the iterative methodology, which uses the previous training results instead of retraining, and its feasibility is strictly verified theoretically. Finally, experiments on bench- mark data sets corroborate a significant saving of the training time with the same number of support vectors and predictive accuracy compared with the original pruning algorithms, and this speedup scheme is also extended to classification problem.展开更多
Hot components operate in a high-temperature and high-pressure environment. The occurrence of a fault in hot components leads to high economic losses. In general, exhaust gas temperature(EGT) is used to monitor the pe...Hot components operate in a high-temperature and high-pressure environment. The occurrence of a fault in hot components leads to high economic losses. In general, exhaust gas temperature(EGT) is used to monitor the performance of hot components.However, during the early stages of a failure, the fault information is weak, and is simultaneously affected by various types of interference, such as the complex working conditions, ambient conditions, gradual performance degradation of the compressors and turbines, and noise. Additionally, inadequate effective information of the gas turbine also restricts the establishment of the detection model. To solve the above problems, this paper proposes an anomaly detection method based on frequent pattern extraction. A frequent pattern model(FPM) is applied to indicate the inherent regularity of change in EGT occurring from different types of interference. In this study, based on a genetic algorithm and support vector machine regression, the relationship model between the EGT and interference was tentatively built. The modeling accuracy was then further improved through the selection of the kernel function and training data. Experiments indicate that the optimal kernel function is linear and that the optimal training data should be balanced in addition to covering the appropriate range of operating conditions and ambient temperature. Furthermore, the thresholds based on the Pauta criterion that is automatically obtained during the modeling process, are used to determine whether hot components are operating abnormally. Moreover, the FPM is compared with the similarity theory, which demonstrates that the FPM can better suppress the effect of the component performance degradation and fuel heat value fluctuation. Finally, the effectiveness of the proposed method is validated on seven months of actual data obtained from a Titan130 gas turbine on an offshore oil platform. The results indicate that the proposed method can sensitively detect malfunctions in hot components during the early stages of a fault, and is robust to various types of interference.展开更多
In this work, two chemometrics methods are applied for the modeling and prediction of electrophoretic mobilities of some organic and inorganic compounds. The successive projection algorithm, feature selection （SPA） ...In this work, two chemometrics methods are applied for the modeling and prediction of electrophoretic mobilities of some organic and inorganic compounds. The successive projection algorithm, feature selection （SPA） strategy, is used as the descriptor selection and model development method. Then, the support vector machine （SVM） and multiple linear regression （MLR） model are utilized to construct the non-linear and linear quantitative structure-property relationship models. The results obtained using the SVM model are compared with those obtained using MLR reveal that the SVM model is of much better predictive value than the MLR one. The root-mean-square errors for the training set and the test set for the SVM model were 0.1911 and 0.2569, respectively, while by the MLR model, they were 0.4908 and 0.6494, respectively. The results show that the SVM model drastically enhances the ability of prediction in QSPR studies and is superior to the MLR model.展开更多
基金supported by the National Natural Science Foundation of China(50576033)
文摘The pruning algorithms for sparse least squares support vector regression machine are common methods, and easily com- prehensible, but the computational burden in the training phase is heavy due to the retraining in performing the pruning process, which is not favorable for their applications. To this end, an im- proved scheme is proposed to accelerate sparse least squares support vector regression machine. A major advantage of this new scheme is based on the iterative methodology, which uses the previous training results instead of retraining, and its feasibility is strictly verified theoretically. Finally, experiments on bench- mark data sets corroborate a significant saving of the training time with the same number of support vectors and predictive accuracy compared with the original pruning algorithms, and this speedup scheme is also extended to classification problem.
文摘Hot components operate in a high-temperature and high-pressure environment. The occurrence of a fault in hot components leads to high economic losses. In general, exhaust gas temperature(EGT) is used to monitor the performance of hot components.However, during the early stages of a failure, the fault information is weak, and is simultaneously affected by various types of interference, such as the complex working conditions, ambient conditions, gradual performance degradation of the compressors and turbines, and noise. Additionally, inadequate effective information of the gas turbine also restricts the establishment of the detection model. To solve the above problems, this paper proposes an anomaly detection method based on frequent pattern extraction. A frequent pattern model(FPM) is applied to indicate the inherent regularity of change in EGT occurring from different types of interference. In this study, based on a genetic algorithm and support vector machine regression, the relationship model between the EGT and interference was tentatively built. The modeling accuracy was then further improved through the selection of the kernel function and training data. Experiments indicate that the optimal kernel function is linear and that the optimal training data should be balanced in addition to covering the appropriate range of operating conditions and ambient temperature. Furthermore, the thresholds based on the Pauta criterion that is automatically obtained during the modeling process, are used to determine whether hot components are operating abnormally. Moreover, the FPM is compared with the similarity theory, which demonstrates that the FPM can better suppress the effect of the component performance degradation and fuel heat value fluctuation. Finally, the effectiveness of the proposed method is validated on seven months of actual data obtained from a Titan130 gas turbine on an offshore oil platform. The results indicate that the proposed method can sensitively detect malfunctions in hot components during the early stages of a fault, and is robust to various types of interference.
文摘In this work, two chemometrics methods are applied for the modeling and prediction of electrophoretic mobilities of some organic and inorganic compounds. The successive projection algorithm, feature selection （SPA） strategy, is used as the descriptor selection and model development method. Then, the support vector machine （SVM） and multiple linear regression （MLR） model are utilized to construct the non-linear and linear quantitative structure-property relationship models. The results obtained using the SVM model are compared with those obtained using MLR reveal that the SVM model is of much better predictive value than the MLR one. The root-mean-square errors for the training set and the test set for the SVM model were 0.1911 and 0.2569, respectively, while by the MLR model, they were 0.4908 and 0.6494, respectively. The results show that the SVM model drastically enhances the ability of prediction in QSPR studies and is superior to the MLR model.
