Applied Mechanics and Materials Vols. 321-324

Abstract: The relationship of mutual information and generalized correlative coefficient is studied, a fuzzy reasoning model based on universal combination operation model is proposed. According to position relation with different parts of medical image and morph feature of the pathological region, a method is proposed based on feature extraction of tree structure considering the adjacent parts of tree structure guided medical prior knowledge. The experimental results show that the proposed algorithm can obtain a higher diagnostic accuracy.

Abstract: In order to improve the safety management of chemical industry park, a novel methodology was probed into from the perspective of safety early-warning by introducing immune mechanism. Based on the corresponding relationship between the immune system and the chemical industry park, the interaction mechanism of antigen and antibody was taken to guide the formulation of the safety early-warning index system. The weights of the indexes were observed and calculated through the improved analytic hierarchy process. Inspired by fuzzy inference model and neural network algorithm, the safety early-warning model and safety forecast model of the chemical industry park were put forward. The application result showed that the methodology can realize the function of safety early-warning and safety forecast for chemical industry park. Furthermore, the corresponding countermeasures can be taken to improve the safety management of the chemical industry park.

Abstract: In the course of the comprehensive diagnosis of equipment troubles, infrared diagnosis technique is adopted to achieve the non-contact NDT (non-destructive testing) and fault diagnosis for the equipments in operation. The oil analysis and integrated monitoring expert system facilitates a comprehensive and systematic monitoring of the lubricating system of important equipments and the conditions of crucial movement parts, providing a development trend of the wearing process of the major friction pairs. It also detects the signs of failure so as to prevent sudden breakdowns, thus enhancing the reliability of condition monitoring and fault diagnosis.

Abstract: In modern industrial enterprises, the level of systematization and automation keeps improving. In light of the complicated conditions that involve multiple factors, multiple failures and multiple processes in the operation of the equipments and by taking fault diagnosis of the unfavorable signs of commutation for DC motors as an example, the paper discusses the principles and application of the mixed integrated intelligent diagnosis system, taking advantage of the knowledge-based expert system and the intelligent diagnosis system of neural network ensemble.

Abstract: Chemical accidents happen in workplaces all around the world. To improve the accident emergency management system is an important methodology to reduce the casualty and loss caused by accidents. In the light of the principles of immune system, the paper tried to analyze the accident emergency management system of chemical industrial park. It elaborated the similarities between immune system and accident emergency management system in the aspects of their existing environment, action object, function and adjustment mechanism. Immune mechanisms were used to formulate the assessment indexes of accident emergency management system performance. Inspired by immune system that keeps the living body healthy in “no antigen condition” and “touching antigen condition”, the assessment index system was formulated from the perspectives of “non-emergency condition” and “emergency condition”. Through the improved analytic hierarchy process, the weights of assessment indexes were calculated. Furthermore, the linear weighted model was applied in the building of the assessment model, which was then used to analyze the accident emergency management system of a chemical industrial park in Guangdong province, China. Through the data from expert investigation questionnaires, the scores of nine assessment indexes reflected accident emergency management system performance were obtained. The result showed that the model is helpful in analyzing the accident emergency management system performance and in improving the accident emergency management system in chemical industrial parks.

Abstract: In order to explore the characteristics of passenger waiting time in high-speed rail hub, this paper analyzed the influencing factors of passenger waiting time, based on the survey of passenger waiting time in high-speed rail hub. And the main influencing factors were screened out using variance analysis. Then the prediction model of passenger waiting time based on BP neural network was established, the parameters of the model were calibrated and the validity was verified. The results show that, travel time in urban area, trip distance, familiarity toward the hub, educational background of passengers, and the type of transportation is the main influencing factor of passenger waiting time in high-speed rail hub, and the average relative error is only 9.2% using the proposed prediction model of passenger waiting time based on BP neural network.

Abstract: To overcome the disadvantages that image analysis of agricultural disease diagnosis was not practical in the field, and the expert diagnosis system had an unsatisfied correct rate, a diagnostic model based on fuzzy rule and BP neural network (back propagation neural network) was constructed. The input vector in the model was formed by a unified description of symptoms using plant protection terms and combined with the membership. The intelligent diagnostic system of vegetable diseases based on the diagnostic model was developed by the mixed programming of Visual C # and Matlab. The test shows that the diagnostic correct rate of the system is 88.95%, and it has better fault tolerance and practical value.

Abstract: The classical multi-class logistic regression classifier uses Newton method to optimize its loss function and suffers the expensive computations and the un-stable iteration process. In our work, we apply the state-of-art optimization techniques BFGS to train multi-class logistic regression and compare them with Newton method on the classification accuracy of 25 datasets experimentally. The results show that BFGS achieves better classification accuracy than the Newton method. Moreover, BFGS have the lower time complexity, in contrast with Newton method. Finally, we also observe that logistic classifier with BFGS method demonstrate comparable performance with the SVM classifier.

Abstract: Least squares support vector machine (LS-SVM) has an outstanding advantage of lower computational complexity than that of standard support vector machines. Its shortcomings are the loss of sparseness and robustness. Thus it usually results in slow testing speed and poor generalization performance. In this paper, a least squares support vector machine with L1 penalty (L1-LS-SVM) is proposed to deal with above shortcomings. A minimum of 1-norm based object function is chosen to get the sparse and robust solution based on the idea of basis pursuit (BP) in the whole feasibility region. Some UCI datasets are used to demonstrate the effectiveness of this model. The experimental results show that L1-LS-SVM can obtain a small number of support vectors and improve the generalization ability of LS-SVM.

Abstract: The methodology is proposed to forecast the daily SSE Composite Index based on artificial neural network and wavelet analysis. The original Composite Index series is decomposed into various components using wavelet techniques at first. The neural network is applied for modeling components of the decomposed series. The final forecast is obtained by combining the components series forecasts. The empirical results show the superior performance of the proposed methodology as compared to the neural network forecasting models. In addition, the results show the obvious difference among different type network in forecasting performance.