Search

This article aims to provide a more practical forest fire spread model for predicting and managing forest fires in Heilongjiang Province, China. Heilongjiang is dominated by spatially heterogeneous combustible forests with undulating terrain and steep slopes. In this article, an artificial neural network framework is used to generate an accurate flame propagation map. Considering inherent fire propagation uncertainties, a fire propagation model containing multidimensional physical and environmental variables is established. Based on fire propagation predictions, the physical fire propagation method is also effectively understood. Additionally, the artificial neural network model can analyse spatial time series patterns and is not a traditional fire spread model. Moreover, this study established a forest fire spread prediction model combining Heilongjiang's cellular automata and Wang Zhengfei's model for comparison with the artificial neural network model. After repeated training and testing, the forest fire prediction results based on the artificial neural network, were found to have average accuracy, sensitivity, and F-measure values of 85.02%, 95.26%, and 89.85%, respectively. The proposed model is suitable for prediction of fire spread beyond large forest fires (more than 1 ha and less than 100 ha of affected forest). Therefore, the model facilitates a better understanding of fire cover propagation behaviours and quickly generates fire peak profiles. The proposed model can enable forest managers and firefighting agencies to plan better firefighting operations and improved firefighting strategy effectiveness.

The traditional modeling methods on suspension force of magnetic bearings are mainly based on static fields analysis, and lack of consideration for the effect of real-time temperature and rotation speed. However, the action area of Maxwell force and the air gap size in the calculation formula of the suspension force dynamically change with temperature rise and rotation speed. Therefore, the accuracy of model results on suspension force is compromised. In this article, taking the centripetal force type-magnetic bearing as an example, the error sources that affect the accuracy of the suspension force are analyzed meticulously. Then, based on the suspension force error sources results, a dynamic model involving multiple dimensional variables is proposed, which includes control current, displacement, temperature, and rotational speed. Superior to the traditional static models, the temperature and rotation speed are integrated into the proposed multidimensional model besides the traditional current and displacement. To prove the validity of suspension force error sources analysis and the accuracy of the proposed model, the dynamic stiffness experiments are carried out. The experiment results show that the multidimensional model is more practical and has higher precision than the traditional static model. [ABSTRACT FROM AUTHOR]

We consider multivariate spacings blocks in the usual sense (refer to Deheuvels (Z Wahrsch Verw Gebiete 64:411-424, 1983)). We consider the sets of exceptional points in the neighborhood of which such spacings are, infinitely often, unusually large. Our main result, in the spirit of Hawkes (Math Proc Camb Phil Soc 293-303, 1981), shows that these sets constitute random fractals, whose Hausdorff dimensions are explicitly evaluated. [ABSTRACT FROM AUTHOR]

• A prediction model of forest fire spread based on artificial neural network is proposed. • Created Wang Zhengfei-CA model suitable for Heilongjiang Forest. • Elevation, temperature, and surface soil moisture content are the most important factors affecting the spread of fire. This article aims to provide a more practical forest fire spread model for predicting and managing forest fires in Heilongjiang Province, China. Heilongjiang is dominated by spatially heterogeneous combustible forests with undulating terrain and steep slopes. In this article, an artificial neural network framework is used to generate an accurate flame propagation map. Considering inherent fire propagation uncertainties, a fire propagation model containing multidimensional physical and environmental variables is established. Based on fire propagation predictions, the physical fire propagation method is also effectively understood. Additionally, the artificial neural network model can analyse spatial time series patterns and is not a traditional fire spread model. Moreover, this study established a forest fire spread prediction model combining Heilongjiang's cellular automata and Wang Zhengfei's model for comparison with the artificial neural network model. After repeated training and testing, the forest fire prediction results based on the artificial neural network, were found to have average accuracy, sensitivity, and F-measure values of 85.02%, 95.26%, and 89.85%, respectively. The proposed model is suitable for prediction of fire spread beyond large forest fires (more than 1 ha and less than 100 ha of affected forest). Therefore, the model facilitates a better understanding of fire cover propagation behaviours and quickly generates fire peak profiles. The proposed model can enable forest managers and firefighting agencies to plan better firefighting operations and improved firefighting strategy effectiveness. [ABSTRACT FROM AUTHOR]

As an emerging technology, artificial intelligence has been applied to identify various physical disorders. Here, we developed a three-layer diagnosis system for lung cancer, in which three machine learning approaches including decision tree C5.0, artificial neural network (ANN) and support vector machine (SVM) were involved. The area under the curve (AUC) was employed to evaluate their decision powers. In the first layer, the AUCs of C5.0, ANN and SVM were 0.676, 0.736 and 0.640, ANN was better than C5.0 and SVM. In the second layer, ANN was similar with SVM but superior to C5.0 supported by the AUCs of 0.804, 0.889 and 0.825. Much higher AUCs of 0.908, 0.910 and 0.849 were identified in the third layer, where the highest sensitivity of 94.12% was found in C5.0. These data proposed a three-layer diagnosis system for lung cancer: ANN was used as a broad-spectrum screening subsystem basing on 14 epidemiological data and clinical symptoms, which was firstly adopted to screen high-risk groups; then, combining with additional 5 tumor biomarkers, ANN was used as an auxiliary diagnosis subsystem to determine the suspected lung cancer patients; C5.0 was finally employed to confirm lung cancer patients basing on 22 CT nodule-based radiomic features.