Applied Mechanics and Materials Vol. 472

Abstract: Due to the lack of the wings detail design progress and the limitation of ordinary detail design ways for the complicated design of wings, while a new wing detail design process was proposed based on traditional wing structural design approach, including two parts: the sub-components design and the particular design. The process involves taking loads on initial proofing structure, structural design, FEM (Finite Element Methods) analysis, and buckling analysis, etc. In the particular design, the structural loads were calculated by the corresponding deformation based on the initial proofing design. The detail components are designed based on the new design process which meets to all the design requirements. It shows that the new design process is feasible and available.

Abstract: In this paper, a method is put forward to forecast the MOVs state based on the improve recursive neural network. The result indicates that recursive network is more adapted to the state forecast of MOV. Because the running state of MOV is closely related to the system voltage and the environment, the state forecast method affected by multi-factors should be further considered.

Abstract: When knocking at crack eggs and nondestructive egg, a flexible piezoelectric film sensor shall be adopted to obtain frequency domain characteristic signals. Consequently, the frequency domain characteristic curves of nondestructive eggs show obvious main frequency domain values, while those of crack eggs show multiple and irregular peak values. In accordance with equal interval frequencies, the first 16 maximum and minimum amplitude values are selected orderly as input vector training Generalized regression neural network .Its Predicted output vector is the egg cracks.. In order to improve prediction accuracy, smooth factor of GRNN neural network is calculated by Particle Swarm Optimization. The experimental result showed that accurate prediction rate of GRNN is up to 94%.

Abstract: In this paper, it was presented that the establishment and experimental investigation of a salt-gradient solar pond. The solar pond was filled with salty water to form three zones (e.g., upper convective zone, non-convective zone and lower convective zone) accordingly with different methods of saline injection. Parameters like salinity and temperature were measured and recorded daily at various locations in the salt-gradient solar pond. The results showed that solar pond collected and stored solar energy for a long period of time can be possible by controlling the thickness and salinity of salt gradient layer of the solar pond.

Abstract: This paper gives the electrical performance characteristics of vacuum glass building integrated photovoltaic (BIPV) modules used as roofing system of building markets. Considering materials and structures are totally different from that of traditional PV module. The optimum power rating condition need be evaluated and analyzed to obtain irradiance and temperature dependence. A temperature and irradiance matrix of performance parameters of a BIPV module is given to predict the energy produced by this BIPV product. To define a suitable standard test condition of vacuum glass BIPV module, the electrical performances under different incident angle and sunlight spectrum are also measured and discussed in this paper.

Abstract: With the development of ubiquitous computing, augmented reality (AR ) technology has become an important research direction. For its Unique properties, it can be used in many engineering areas, especially in biomedical engineering. In this article, we mainly summarized the technical feature and technological superiority of the AR technology, and gave some applications and effectiveness of AR-based biomedical Engineering. This will be helpful for those who are eager to use this brand new technology in medicine.

Abstract: This paper proposes a novel multi-radius density clustering algorithm based on outlier factor. The algorithm first calculates the density-similar-neighbor-based outlier factor (DSNOF) for each point in the dataset according to the relationship of the density of the point and its neighbors, and then treats the point whose DSNOF is smaller than 1 as a core point. Second, the core points are used for clustering by the similar process of the density based spatial clustering application with noise (DBSCAN) to get some sub-clusters. Third, the proposed algorithm merges the obtained sub-clusters into some clusters. Finally, the points whose DSNOF are larger than 1 are assigned into these clusters. Experiments are performed on some real datasets of the UCI Machine Learning Repository and the experiments results verify that the effectiveness of the proposed model is higher than the DBSCAN algorithm and k-means algorithm and would not be affected by the parameter greatly.

Abstract: Credit risk assessment is critical factor in credit risk management, which has played a key role in financial and banking industry. Many classification methods are used in credit risk assessment aiming to establish classifiers to predict the credit state of the corporate (good or bad). However, most of classification methods can not handle continuous variables. So, continuous variables must be quantified. In this paper, we first propose an improved quantization method, namely IDM, based on the statistical independence; then we use data mining techniques, i.e., C4.5 decision tree, Naive-Bayes and SVM classifier, to classify and predict the quantified credit data. The aim is to investigate the effect of quantization method on the classification of credit approval data. The Experimental results show that our approach significantly improves the mean accuracy of classification than other known quantization methods. This denotes that the proposed method can make an effective interpretation and point out the ability of design of a new intelligent assistance credit approval data system.

Abstract: This study described a five-factor typology of online consumers motivations labeled as the functional shopper, the following shopper, the surfing shopper, conflicting shopper and e-laggard, and a five-factor typology of the clickstream data labeled as the functional browser, the hedonic browser, the impulsive browser, the comparative browser and the knowledge building browser by cluster analysis. A correspondence analysis for two typologies demonstrated an independent cluster frame with limited correspondence. Theoretical and marketing implications are discussed.

Abstract: Recommender system (RS) has been evaluated in many but incomparable ways beyond accuracy and thus proposing an evaluation framework to synthesize the existing strategies seems a solution. However, few scholars did it so far. Through literature review, user interview and expert assessment, this study proposed a theoretical evaluation model of RS and then formed the assessment tool, RS Evaluation Questionnaire (RSE). The results showed that RSE was an effective tool to evaluate a recommender system, with its reliability (Cronbachs α=0.803) and validity meeting the requirements of psychometrics. Seven factors such as Perceived Quality and Perceived Ease of Use were generated by factor analysis, accounting for 63.126% of the variance. Furthermore, regression analysis indicated that different combinations of RSE factors could significantly predict User Satisfaction, Reuse Intention and positive Word-Of-Mouth (WOM) spreading willingness. Enlightenments for future research and practice were discussed as well in the end.