Applied Mechanics and Materials Vols. 333-335

Abstract: In the era of knowledge economy, knowledge has become the most critical enterprise resources; however, because of the accelerating speed of knowledge innovation, knowledge sharing between organizations becomes the inevitable choice for enterprises. Moreover, selection of the appropriate sharing partners has an important impact on the efficiency of knowledge sharing. In this paper, the authors analyze the factors of knowledge sharing among the cluster enterprises, build a cluster enterprise knowledge-sharing partner selection index system, and use the BP neural network model to select suitable enterprise knowledge sharing partners. Finally, the authors demonstrate the feasibility of the method with an example.

Abstract: Clustering algorithms, like K-means Algorithm, use distances in attribute space to cluster data. However the computation of distances in attribute space influences the accuracy. So innovatively, Variance-Similarity clustering algorithm defines similarity as a function of the attribute variance, and clusters data by the comparison of similarities. In computer simulation, the comparison of Variance-Similarity Algorithm and K-means Algorithm on UCI data sets presents that Variance-Similarity Algorithm has a better clustering accuracy than K-means Algorithm.

Abstract: Resource scheduling is still a challenge in organizational design. In order to overcome it, organizational design flow is analyzed in detail. A model is presented in the paper according to the three stages organizational design flow. In order to optimize resource scheduling, a kind of ant colony algorithm is proposed, which has laid a solid foundation for optimization of organizational design.

Abstract: The traditional approach to deal with incomplete information system is to make it completed, when a new object added only need a static attribute reduction algorithm to obtain the rules, wastes a lot of resources. The goal of incremental rules mining is to maintain the consistency of the rules in incomplete decision table. When a new object is added, establish discernibility matrix of the original decision table, get distribution reduction set, then generate conjunctive items export rules set. It introduces incremental learning concept, avoids tedious counting process. It can be effective for large-scale incomplete ocean data reduction and it also provides a strong basis for decision making for the marine environment processing and follow-up processing.

Abstract: This paper points out the bottleneck of classical Apriori algorithm, presents an improved association rule mining algorithm based on Apriori algorithm.The new algorithm is based on pruing away the itemsets whose support degree is less than minsupport to reduce the number of itemsets in the transaction database. At the same time the new algorithm change the candidate_gen function to generate a continuous access page. According to the running result of the algorithm, the processing time of mining is decreased and the efficiency of algorithm has improved.Whats more, the new algorithm can find the learners frequent traversal path to improve the intelligence of the distance education platform. Keywords: Associaion Rules;Apriori Algorithm; Frequent Traversal Path;Distance Education Platform

Abstract: Emergency management is such a domain where experiential knowledge could be easily collected, and is quite suitable for the application of case based reasoning. However, in practice there are two problems limiting the effectiveness of CBR, the he incomplete information and changing situations. This paper proposed an approach based on fuzzy sets and text mining to solve those two problems, which contains four steps: a) represent the attributes with fuzzy sets, b) extract solution texts with text classification, c) establish connections of attributes and solutions with association rules, and d) adjust the solution with fuzzy reasoning. An example shows the adaption for emergency management and illustrates the improvement for CBR with the approach.

Abstract: To investigate the different modes of human thinking, we designed an eye tracking experiment during people recognized two category images of histograms and scenes, and used the support vector machine (SVM) classification algorithm to classify these eye movement data. The results of statistical analysis showed that there were significant differences in saccade distance and pupil diameter between these two category images. By the feature selection, normalization of data preprocessing, and SVM classification, the results of classification analysis showed that there was a better performance on the classification of the histograms and scenes. These results suggest we can identify the modes of human thinking through the SVM classification methods based on the eye movement data.

Abstract: An intelligent quality diagnosis method for process quality diagnosis and improvement was proposed to find out influencing input parameters for output quality and then provide suggestion for quality engineering to adjust them to acquire high quality performance. The diagnosis method extends the traditional quality control and diagnosis method that only for the output quality of manufacturing process. It can detect the input parameters of the manufacturing process and provide sensitivities of input parameter for adjustment. BN-MTY method was applied to explain the reason of quality failure in T² control chart and the root output quality indicators that aroused the process quality anomaly was located. The integrated method of neural network and sensitivity analysis was applied to get the weight and threshold value of neural cell in the forecasting network. his integrated quality diagnosis method can diagnose the input parameters and provide accurate sensitivities for quality improvement.

Abstract: This paper proposed a cooperative receding horizon optimal control framework, based on differential flatness and B-splines, which was used to solve the real-time cooperative trajectory planning for multi-UCAV performing cooperative air-to-ground target attack missions. The planning problem was formulated as a cooperative receding horizon optimal control problem (CRHC-OCP), and then the differential flatness and B-splines were introduced to lower the dimension of the planning space and parameterize the spatial trajectories. Moreover, for the dynamic and uncertainty of the battlefield environment, the cooperative receding horizon control was introduced. Finally, the proposed approach is demonstrated, and the results show that this approach is feasible and effective.

Abstract: Support Vector Machine (SVM) is one of the most popular and effective data mining algorithms which can be used to resolve classification or regression problems, and has attracted much attention these years. SVM could find the optimal separating hyperplane between classes, which afford outstanding generalization ability with it. Usually all the labeled records are used as training set. However, the optimal separating hyperplane only depends on a few crucial samples (Support Vectors, SVs), we neednt train SVM model on the whole training set. In this paper a novel SVM model based on K-means clustering is presented, in which only a small subset of the original training set is selected to constitute the final training set, and the SVM classifier is built through training on these selected samples. This greatly decrease the scale of the training set, and effectively saves the training and predicting cost of SVM, meanwhile guarantees its generalization performance.