Least squares weighted twin support vector machines with local information

A least squares version of the recently proposed weighted twin support vector machine with local information(WLTSVM) for binary classification is formulated. This formulation leads to an extremely simple and fast algorithm, called least squares weighted twin support vector machine with local information(LSWLTSVM), for generating binary classifiers based on two non-parallel hyperplanes. Two modified primal problems of WLTSVM are attempted to solve, instead of two dual problems usually solved. The solution of the two modified problems reduces to solving just two systems of linear equations as opposed to solving two quadratic programming problems along with two systems of linear equations in WLTSVM. Moreover, two extra modifications were proposed in LSWLTSVM to improve the generalization capability. One is that a hot kernel function, not the simple-minded definition in WLTSVM, is used to define the weight matrix of adjacency graph, which ensures that the underlying similarity information between any pair of data points in the same class can be fully reflected. The other is that the weight for each point in the contrary class is considered in constructing equality constraints, which makes LSWLTSVM less sensitive to noise points than WLTSVM. Experimental results indicate that LSWLTSVM has comparable classification accuracy to that of WLTSVM but with remarkably less computational time.

A review of real-time multi-GNSS precise orbit determination based on the filter method

Stable and reliable high-precision satellite orbit products are the prerequisites for the positioning services with high performance.In general,the positioning accuracy depends strongly on the quality of satellite orbit and clock products,especially for absolute positioning modes,such as Precise Point Positioning(PPP).With the development of real-time services,real-time Precise Orbit Determination(POD)is indispensable and mainly includes two methods:the ultra-rapid orbit prediction and the real-time filtering orbit determination.The real-time filtering method has a great potential to obtain more stable and reliable products than the ultra-rapid orbit prediction method and thus has attracted increasing attention in commercial companies and research institutes.However,several key issues should be resolved,including the refinement of satellite dynamic stochastic models,adaptive filtering for irregular satellite motions,rapid convergence,and real-time Ambiguity Resolution(AR).This paper reviews and summarizes the current research progress in real-time filtering POD with a focus on the aforementioned issues.In addition,the real-time filtering orbit determination software developed by our group is introduced,and some of the latest results are evaluated.The Three-Dimensional(3D)real-time orbit accuracy of GPS and Galileo satellites is better than 5 cm with AR.In terms of the convergence time and accuracy of kinematic PPP AR,the better performance of the filter orbit products is validated compared to the ultra-rapid orbit products.

Evaluation of Different Predictor Models for Detailed Soil Particle-Size Distribution

An accurate mathematical representation of soil particle-size distribution(PSD) is required to estimate soil hydraulic properties or to compare texture measurements using different classification systems. However, many databases do not contain full PSD data,but instead contain only the clay, silt, and sand mass fractions. The objective of this study was to evaluate the abilities of four PSD models(the Skaggs model, the Fooladmand model, the modified Gray model GM(1,1), and the Fredlund model) to predict detailed PSD using limited soil textural data and to determine the effects of soil texture on the performance of the individual PSD model.The mean absolute error(MAE) and root mean square error(RMSE) were used to measure the goodness-of-fit of the models, and the Akaike's information criterion(AIC) was used to compare the quality of model fits. The performance of all PSD models except the GM(1,1) improved with increasing clay content in soils. This result showed that the GM(1,1) was less dependent on soil texture.The Fredlund model was the best for describing the PSDs of all soil textures except in the sand textural class. However, the GM(1,1) showed better performance as the sand content increased. These results indicated that the Fredlund model showed the best performance and the least values of all evaluation criteria, and can be used using limited soil textural data for detailed PSD.