Multi-label dimensionality reduction and classification with extreme learning machines

In the need of some real applications, such as text categorization and image classification, the multi-label learning gradually becomes a hot research point in recent years. Much attention has been paid to the research of multi-label classification algorithms. Considering the fact that the high dimensionality of the multi-label datasets may cause the curse of dimensionality and wil hamper the classification process, a dimensionality reduction algorithm, named multi-label kernel discriminant analysis （MLKDA）, is proposed to reduce the dimensionality of multi-label datasets. MLKDA, with the kernel trick, processes the multi-label integrally and realizes the nonlinear dimensionality reduction with the idea similar with linear discriminant analysis （LDA）. In the classification process of multi-label data, the extreme learning machine （ELM） is an efficient algorithm in the premise of good accuracy. MLKDA, combined with ELM, shows a good performance in multi-label learning experiments with several datasets. The experiments on both static data and data stream show that MLKDA outperforms multi-label dimensionality reduction via dependence maximization （MDDM） and multi-label linear discriminant analysis （MLDA） in cases of balanced datasets and stronger correlation between tags, and ELM is also a good choice for multi-label classification.

Constrained multi-degree reduction of rational Bézier curves using reparameterization

Applying homogeneous coordinates, we extend a newly appeared algorithm of best constrained multi-degree reduc- tion for polynomial Bézier curves to the algorithms of constrained multi-degree reduction for rational Bézier curves. The idea is introducing two criteria, variance criterion and ratio criterion, for reparameterization of rational Bézier curves, which are used to make uniform the weights of the rational Bézier curves as accordant as possible, and then do multi-degree reduction for each component in homogeneous coordinates. Compared with the two traditional algorithms of "cancelling the best linear common divisor" and "shifted Chebyshev polynomial", the two new algorithms presented here using reparameterization have advantages of simplicity and fast computing, being able to preserve high degrees continuity at the end points of the curves, do multi-degree reduction at one time, and have good approximating effect.

Nonlinear consensus protocols for multi-agent systems based on centre manifold reduction

Nonlinear consensus protocols for dynamic directed networks of multi-agent systems with fixed and switching topologies are investigated separately in this paper. Based on the centre manifold reduction technique, nonlinear consensus protocols are presented. We prove that a group of agents can reach a β-consensus, the value of which is the group decision value varying from the minimum and the maximum values of the initial states of the agents. Moreover, we derive the conditions to guarantee that all the agents reach a β-consensus on a desired group decision value. Finally, a simulation study concerning the vertical alignment manoeuvere of a team of unmanned air vehicles is performed. Simulation results show that the nonlinear consensus protocols proposed are more effective than the linear protocols for the formation control of the agents and they are an improvement over existing protocols.

Electrochemical Reduction of Oxygen on Multi-walled Carbon Nanotubes Electrode in Alkaline Solution

The multi-walled carbon nanotubes (MWNTs) electrode was constructed using poly-tetrafluoroethylene as binder, and the electrochemical reductive behavior of oxygen in alkaline solution was first examined on this electrode. Compared with other carbon materials, MWNTs show higher electrocatalytic activity, and the reversibility of O2 reduction reaction is greatly improved. The experiments reveal that the electrochemical reduction of O2 to HO2- is controlled by adsorption. The preliminary results illustrate the potential application of MWNTs in fuel cells.

Similarity Reductions of(2+1)-Dimensional Multi-component Broer-Kaup System

Painlevé property of the (2+1)-dimensional multi-component Broer-Kaup (BK)system is considered byusing the standard Weiss-Kruskal approaches.Applying the Clarkson and Kruskal (CK) direct method to the (2+1)-dimensional multi-component BK system,some types of similarity reductions are obtained.By solving the reductions,one can get the solutions of the (2+1)-dimensional multi-component BK system.

STUDY ON SPECKLE REDUCTION IN MULTI-LOOK POLARIMETRIC SAR IMAGE

This paper studies the speckle reduction in multi-look polarimetric synthetic aperture radar (SAR) image. A multi-look polarimetric whitening filtering (MPWF) method is presented and extended to form a fully polarimetric filter with multi-channel output. The paper also quantifies the speckle reduction amount achievable by the MPWF, and compares the MPWF with the span, weighting and power equalization methods. Experimental results with the NASA/JPL L-band 4-look polarimetric SAR data verify the effectiveness and superiority of the MPWF, and show that the MPWF is of great advantage for enhancing SAR image classification.

Estimating the Macroeconomic Costs of CO_2 Emission Reduction in China Based on Multi-objective Programming

This paper estimates the macroeconomic costs of CO_2 emission reduction in China employing the input-output analysis with the multi-objective programming approach.The results show that the effect of reducing CO_2 emissions on China's economy is significant.Under the present conditions,the estimated macroeconomic costs of CO_2 emission reduction in 2010 for China are approximately 3,100-4,024 RMB t^(-1).The stronger the abatement actions,the higher the macroeconomic costs of per unit emission reduction would be.Excavation industry,oil industry,chemical industry,and metal smelting industry have high potential to abate their CO_2 emissions.

2-Tuple and Rough Set Based Reduction Model for Multi-sensory Evaluation Indicators

In order to lessen adverse influences of excessive evaluative indicators of the initial set in multi-sensory evaluation,a2-tuple and rough set based reduction model is built to simplify the initial set of evaluative indicators. In the model,a great variety of descriptive forms of the multi-sensory evaluation are also taken into consideration. As a result,the method proves effective in reducing redundant indexes and minimizing index overlaps without compromising the integrity of the evaluation system. By applying the model in a multi-sensory evaluation involving community public information service facilities,the research shows that the results are satisfactory when using genetic algorithm optimized BP neural network as a calculation tool. It shows that using the reduced and simplified set of indicators has a better predication performance than the initial set,and 2-tuple and rough set based model offers an efficient way to reduce indicator redundancy and improves prediction capability of the evaluation model.

Influence of a multi-step process on the thickness reduction error of sheet metal in a flexible rolling process

Flexible rolling is a forming process based on thickness reduction, and the precision of thickness reduction is the key factor affecting bending deformation. The major purpose of the present work is to solve the problem of bending deformation error caused by insufficient thickness reduction. Under the condition of different rolling reductions with the same sheet thickness and the same thickness reduction with different sheet thicknesses, the thickness reduction error of sheet metal is analyzed. In addition, the bending deformation of sheet metal under the same conditions is discussed and the influence of the multi-step forming process on the thickness reduction error is studied. The results show that, under the condition of the same sheet thickness, the thickness reduction error increases with increasing rolling reduction because of an increase in work hardening. As rolling reduction increases, the longitudinal bending deformation decreases because of the decrease of the maximum thickness difference. Under the condition with the same thickness reduction, the thickness reduction error increases because of the decrease of the rolling force with increasing sheet thickness. As the sheet thickness increases, the longitudinal bending deformation increases because of the increase in the maximum thickness difference. A larger bending deformation is divided into a number of small bending deformations in a multi-step forming process, avoiding a sharp increase in the degree of work hardening; the thickness reduction error is effectively reduced in the multi-step forming process. Numerical simulation results agree with the results of the forming experiments.

Multi-label dimensionality reduction based on semi-supervised discriminant analysis

Multi-label data with high dimensionality often occurs,which will produce large time and energy overheads when directly used in classification tasks.To solve this problem,a novel algorithm called multi-label dimensionality reduction via semi-supervised discriminant analysis(MSDA) was proposed.It was expected to derive an objective discriminant function as smooth as possible on the data manifold by multi-label learning and semi-supervised learning.By virtue of the latent imformation,which was provided by the graph weighted matrix of sample attributes and the similarity correlation matrix of partial sample labels,MSDA readily made the separability between different classes achieve maximization and estimated the intrinsic geometric structure in the lower manifold space by employing unlabeled data.Extensive experimental results on several real multi-label datasets show that after dimensionality reduction using MSDA,the average classification accuracy is about 9.71% higher than that of other algorithms,and several evaluation metrices like Hamming-loss are also superior to those of other dimensionality reduction methods.

A CNN-Based Single-Stage Occlusion Real-Time Target Detection Method

Aiming at the problem of low accuracy of traditional target detection methods for target detection in endoscopes in substation environments, a CNN-based real-time detection method for masked targets is proposed. The method adopts the overall design of backbone network, detection network and algorithmic parameter optimisation method, completes the model training on the self-constructed occlusion target dataset, and adopts the multi-scale perception method for target detection. The HNM algorithm is used to screen positive and negative samples during the training process, and the NMS algorithm is used to post-process the prediction results during the detection process to improve the detection efficiency. After experimental validation, the obtained model has the multi-class average predicted value (mAP) of the dataset. It has general advantages over traditional target detection methods. The detection time of a single target on FDDB dataset is 39 ms, which can meet the need of real-time target detection. In addition, the project team has successfully deployed the method into substations and put it into use in many places in Beijing, which is important for achieving the anomaly of occlusion target detection.

基于Multi-Stage DEA模型的中国细羊毛生产技术效率实证分析

以内蒙古等4省区为例，采用基于规模报酬可变假设的产出主导型Muhi-StageDEA模型，对2001～2010年中国细羊毛生产技术效率进行了测算和分析。研究结果表明，中国细羊毛生产技术效率不高，存在显著的技术效率损失；细羊毛生产在总体上出现了纯技术效率下降和规模效率提高并存的现象，2008年以来纯技术效率提高对细羊毛生产技术效率改善的促进作用逐渐增强，而细羊毛生产规模效率则呈现出了一定的下降趋势；各省区细羊毛生产技术效率存在较大差异，细羊毛生产的规模效率均要高于纯技术效率，规模效率仍是影响各省区细羊毛生产技术效率改善和提高的主要因素。最后提出改善和提高中国细羊毛生产技术效率的政策建议。

Noise Reduction of an Axial Piston Pump by Valve Plate Optimization

Current researches mainly focus on the investigations of the valve plate utilizing pressure relief grooves. However,air?release and cavitation can occur near the grooves. The valve plate utilizing damping holes show excellent perfor?mance in avoiding air?release and cavitation. This study aims to reduce the noise emitted from an axial piston pump using a novel valve plate utilizing damping holes. A dynamic pump model is developed,in which the fluid properties are carefully modeled to capture the phenomena of air release and cavitation. The causes of di erent noise sources are investigated using the model. A comprehensive parametric analysis is conducted to enhance the understanding of the e ects of the valve plate parameters on the noise sources. A multi?objective genetic algorithm optimization method is proposed to optimize the parameters of valve plate. The amplitudes of the swash plate moment and flow rates in the inlet and outlet ports are defined as the objective functions. The pressure overshoot and undershoot in the piston chamber are limited by properly constraining the highest and lowest pressure values. A comparison of the various noise sources between the original and optimized designs over a wide range of pressure levels shows that the noise sources are reduced at high pressures. The results of the sound pressure level measurements show that the optimized valve plate reduces the noise level by 1.6 d B(A) at the rated working condition. The proposed method is e ective in reducing the noise of axial piston pumps and contributes to the development of quieter axial piston machines.

Multi-objective optimization for leaching process using improved two-stage guide PSO algorithm

A mathematical mechanism model was proposed for the description and analysis of the heat-stirring-acid leaching process.The model is proved to be effective by experiment.Afterwards,the leaching problem was formulated as a constrained multi-objective optimization problem based on the mechanism model.A two-stage guide multi-objective particle swarm optimization(TSG-MOPSO) algorithm was proposed to solve this optimization problem,which can accelerate the convergence and guarantee the diversity of pareto-optimal front set as well.Computational experiment was conducted to compare the solution by the proposed algorithm with SIGMA-MOPSO by solving the model and with the manual solution in practice.The results indicate that the proposed algorithm shows better performance than SIGMA-MOPSO,and can improve the current manual solutions significantly.The improvements of production time and economic benefit compared with manual solutions are 10.5% and 7.3%,respectively.

Multi-stage ponds-wetlands ecosystem for effective wastewater treatment

The performance of the Dongying multi-stage ponds-wetlands ecosystem was investigated in this work. Study of the removal of different pollutants (BOD5, COD, SS, TP, TN, NH3-N, etc.) in different temperature seasons and different units in this system indicated that effluent BOD5 and SS were constant to less than 11 mg/L and 14 mg/L throughout the experimental proc- esses; but that the removal efficiencies of pollutants such as TP, TN, NH3-N, COD varied greatly with season. The higher the temperature was, the higher was the observed removal in this system. Additionally, each unit of the system functioned differently in removing pollutants. BOD5 and SS were mainly removed in the first three units (hybrid facultative ponds, aeration ponds and aerated fish ponds), whereas nitrogen and phosphates were mainly removed in hydrophyte ponds and constructed reed wetlands. The multi-stage ponds-wetlands ecosystem exhibits good potential of removing different pollutants, and the effluent quality meet several standards for wastewater reuse.

Process design for multi-stage stretch forming of aluminium alloy aircraft skin

A process design approach for multi-stage stretch forming was proposed by combining the strain distribution method and finite element method(FEM)to determine the minimum stage number and deformation amount of each stage.The strain distribution method was used to calculate the deformation amount of each stage and evaluate the formability through a safety criterion.FE simulation was taken as an analysis tool to reveal the deformation behaviour,to predict the strain contour and to determine the process parameters at each stage.To evaluate the effect of heat treatment after pre-strain on occurrence of deformation defects during the subsequent deformation,a multi-stage uniaxial tension test for 2B06 aluminium alloy sheet was carried out.A case study demonstrates that the approach has high reliability and good practicability.

Springback analysis and strategy for multi-stage thin-walled parts with complex geometries

Springback of a SUS321 complex geometry part formed by the multi-stage rigid-flexible compound process was studied through numerical simulations and laboratory experiments in this work.The sensitivity analysis was provided to have an insight in the effect of the evaluated process parameters.Furthermore,in order to minimize the springback problem,an accurate springback simulation model of the part was established and validated.The effects of the element size and timesteps on springback model were further investigated.Results indicate that the custom mesh size is beneficial for the springback simulation,and the four timesteps are found suited for the springback analysis for the complex geometry part.Finally,a strategy for reducing the springback by changing the geometry of the blank is proposed.The optimal blank geometry is obtained and used for manufacturing the part.

Sequential maneuvering decisions based on multi-stage influence diagram in air combat

A multi-stage influence diagram is used to model the pilot＇s sequential decision making in one on one air combat. The model based on the multi-stage influence diagram graphically describes the elements of decision process, and contains a point-mass model for the dynamics of an aircraft and takes into account the decision maker＇s preferences under uncertain conditions. Considering an active opponent, the opponent＇s maneuvers can be modeled stochastically. The solution of multistage influence diagram can be obtained by converting the multistage influence diagram into a two-level optimization problem. The simulation results show the model is effective.

Stress redistribution in multi-stage hydraulic fracturing of horizontal wells in shales

Multi-stage hydraulic fracturing of horizontal wells is the main stimulation method in recovering gas from tight shale gas reservoirs, and stage spacing deter- mination is one of the key issues in fracturing design. The initiation and propagation of hydraulic fractures will cause stress redistribution and may activate natural fractures in the reservoir. Due to the limitation of the analytical method in calculation of induced stresses, we propose a numerical method, which incorporates the interaction of hydraulic fractures and the wellbore, and analyzes the stress distri- bution in the reservoir under different stage spacing. Simulation results indicate the following： （1） The induced stress was overestimated from the analytical method because it did not take into account the interaction between hydraulic fractures and the horizontal wellbore. （2） The hydraulic fracture had a considerable effect on the redis- tribution of stresses in the direction of the horizontal wellbore in the reservoir. The stress in the direction per- pendicular to the horizontal wellbore after hydraulic frac- turing had a minor change compared with the original in situ stress. （3） Stress interferences among fractures were greatly connected with the stage spacing and the distance from the wellbore. When the fracture length was 200 m, and the stage spacing was 50 m, the stress redistribution due to stage fracturing may divert the original stress pat- tern, which might activate natural fractures so as to generate a complex fracture network.