Applied Mechanics and Materials Vols. 411-414

Abstract: The knowledge sharing is one of the most important characteristics of knowledge management. In the traditional model of knowledge management, employees only select the sharing knowledge through independent action, and operating behavior between employees of the same type did not reflect reference. This paper is the integration of the recommendation algorithm of data mining and the traditional knowledge ontology knowledge management model, proposing the process enterprise knowledge management model based on the recommendation algorithm, and knowledge management framework of knowledge as the main body, the field of process-driven and recommendation process for the behavior. To recommend the appropriate knowledge for the staff improves the efficiency of enterprise employees staff to the knowledge and promote the application and innovation of knowledge of the enterprise.

Abstract: The traditional cluster analysis method based on the true distance is not conducive to the accurate calculation of earthquake different fault rupture propagation and healing rate. This paper proposed and gave a new clustering method based on soft distance calculations. The clustering process based on soft distance calculations, the calculation method for soft distances and the specific clustering algorithm based on soft distances are given. For the real sample points of strong earthquake as a data source, we use this clustering method and other traditional clustering methods to cluster and analysis the data source, and analysis results have showed that the clustering method obtained the same cluster center with the earth stress field evolution, so this method has objective truth. The cluster analysis method for the earthquake fault zones in the accurate calculation of the next strong earthquake provides a good basis for the calculations.

Abstract: Partially missing data sets are a prevailing problem in pattern recognition. In this paper, the problem of clustering incomplete data sets is considered, and missing attribute values are imputed by the centers of corresponding nearest-neighbor intervals. Firstly, the algorithm estimates the nearest-neighbor intervals of missing attribute values by using the attribute distribution information of the data sets sufficiently. Secondly, the missing attribute values are imputed by the center of the intervals so as to clustering incomplete data sets. The proposed algorithm introduces the nearest neighbor information into incomplete data clustering, and the comparisons of the experimental results for two UCI data sets demonstrate the capability of the proposed algorithm.

Abstract: This paper combines domain ontology and LDA model to propose a new method of hierarchical web text classification. Experimental results show that the method has good performance with high recall rate and accuracy rate.

Abstract: Seeking a way non-invasive and adaptive to differentiate the normal and abnormal heart sound signals to provide more valuable reference method to the clinical diagnosis. This paper made the largest Lyapunov exponent as the mainline. According to the unity of the whole signal in different stages, a method to study the characteristic in stage was proposed. First of all, we made phase space reconstitution to the typical seven normal and abnormal heart sound signals. Then, we calculated the largest Lyapunov exponents according to the phase space reconstitution parameters. At last, we compared and analyzed the mean values of the largest Lyapunov exponents. The mean value of the normal heart sound signal in S1 was 0.145, which was much larger than that of the abnormal signals and the mean value of the normal heart sound signal in S2 was larger than that of the abnormal ones, too. This conclusion means that there are chaotic characteristic in the heart sound signals and the degree of chaos in normal heart sounds is higher than that in the abnormal heart sound signals.

Abstract: The Independent Component Analysis (ICA) is a classical algorithm for exploring statistically independent non-Gaussian signals from multi-dimensional data, which has a wide range of applications in engineering, for instance, the blind source separation. The classical ICA measures the Gaussian characteristic by kurtosis, which has the following two disadvantages. Firstly, the kurtosis relies on the value of samples, and is not robust to outliers. Secondly, the algorithm often falls into local optima. To address these drawbacks, we replace the kurtosis by negative entropy, utilize the simulated annealing algorithm for optimization, and finally propose an improved ICA algorithm. Experimental results demonstrate that the proposed algorithm outperforms the classical ICA in its robustness to outliers and convergent rate.

Abstract: It is often the case that managers and social scientists are called to deal with time series. Time series analysis usually involves a study of the components of the time series and finding models that permit statistical inferences and predictions. ARIMA models are, in theory, the most general class of models for forecasting a time series. The commonly known Box-Jenkins approach to ARIMA model building is an iterative process. To facilitate the iterative process and to relieve the boredom of computational errands, we have developed an assistor for building ARIMA models. The assistor is implemented in Java with embedded R for statistical functions. With the help of the assistor ARIMA models for time series are few clicks away, thus enabling users to focus their efforts on the decision problems at hand.

Abstract: Accurate modeling of Electroencephalography (EEG) signals is an important problem in clinical diagnosis of brain diseases. The method using support vectors machine (SVM) based on the structure risk minimization provides us an effective way of learning machine. But solving the quadratic programming problem for training SVM becomes a bottle-neck of using SVM because of the long time of SVM training. In this paper, a local-SVM method is proposed for modeling EEG signals. The local method is presented for improving the speed of the prediction of EEG signals. Furthermore, this proposed model is used to detect epilepsy from EEG signals in which dynamical characteristics are difference between normal and epilepsy EEG signals. The experimental results show that the training of the local-SVM obtains a good behavior. In addition, the local SVM method significantly improves the prediction and detection precision.

Abstract: Fiber signal denoising technology is one of the key technologies the OTDR or distributed fiber optic sensors, fiber optic equipment. Actual project application, proposed an improved undecimated discrete wavelet transform (UDWT) denoising technology, relative to traditional wavelet transform denoising technology, has the following characteristics: Denoised curve is more smooth; Better peak detection capability; Better small attenuation maintain capability; Better denoising capability. This technology has achieved very good results in OTDR equipment, beside it can be applied to other optical signal processing.

Abstract: With the development of electronic technology, electronic equipments are suffering more electromagnetic interferences, so it is necessary to detect and evaluate their electromagnetic interference degree. The most commonly used method is the anechoic chamber test method that conducted in closed electromagnetic wave environment. However, there are some disadvantages, such as low detection efficiency and high testing cost. In order to avoid these advantages, a new method based on frequency domain analysis is presented to adaptively identify RF band of a detected device working in a non-Gaussian background environment. In this method, a non-parametric hypothesis testing is used to solve the interference of background noise on the detection result in an open environment. Further, an adaptive threshold hypothesis test method is also used to obtain the radio frequency spectrum of the detected signals for low SNR in order to overcome the frequency domain time-varying characteristics. The simulations show that the adaptive RF identification based on frequency domain analysis can improve the detection accuracy in the non-Gaussian background environment.