Model-Free Feature Screening via Maximal Information Coefficient (MIC) for Ultrahigh-Dimensional Multiclass Classification

It is common for datasets to contain both categorical and continuous variables. However, many feature screening methods designed for high-dimensional classification assume that the variables are continuous. This limits the applicability of existing methods in handling this complex scenario. To address this issue, we propose a model-free feature screening approach for ultra-high-dimensional multi-classification that can handle both categorical and continuous variables. Our proposed feature screening method utilizes the Maximal Information Coefficient to assess the predictive power of the variables. By satisfying certain regularity conditions, we have proven that our screening procedure possesses the sure screening property and ranking consistency properties. To validate the effectiveness of our approach, we conduct simulation studies and provide real data analysis examples to demonstrate its performance in finite samples. In summary, our proposed method offers a solution for effectively screening features in ultra-high-dimensional datasets with a mixture of categorical and continuous covariates.

基于一种距离相关的超高维生存数据Model-Free特征筛选

随着大数据时代的来临,数据维度爆炸式增长,超高维数据的降维问题逐渐成为众多研究领域的热点话题。由于响应变量通常存在右删失,处理超高维完全数据的降维方法在右删失数据中将不再适用。本研究提出一种新的基于距离相关能有效处理超高维右删失数据的特征筛选方法。首先利用距离相关系数计算每个协变量对响应变量的边际效应,建立与该系数有关的筛选指标,然后再根据事先确立的筛选准则进行特征筛选。提出的特征筛选方法不依赖任何模型结构假定,因此可以有效避免模型指定错误带来的不良后果。此外,该方法采用的距离协方差估计量是总体距离协方差的一个无偏估计,统计准确性和计算精度高。模拟和实证研究表明,提出的方法能在保留所有重要变量的前提下快速剔除与响应变量相关程度较弱的协变量,从而达到降低参数维数的目的。

Model-Free Ultra-High-Dimensional Feature Screening for Multi-Classified Response Data Based on Weighted Jensen-Shannon Divergence

In ultra-high-dimensional data, it is common for the response variable to be multi-classified. Therefore, this paper proposes a model-free screening method for variables whose response variable is multi-classified from the point of view of introducing Jensen-Shannon divergence to measure the importance of covariates. The idea of the method is to calculate the Jensen-Shannon divergence between the conditional probability distribution of the covariates on a given response variable and the unconditional probability distribution of the covariates, and then use the probabilities of the response variables as weights to calculate the weighted Jensen-Shannon divergence, where a larger weighted Jensen-Shannon divergence means that the covariates are more important. Additionally, we also investigated an adapted version of the method, which is to measure the relationship between the covariates and the response variable using the weighted Jensen-Shannon divergence adjusted by the logarithmic factor of the number of categories when the number of categories in each covariate varies. Then, through both theoretical and simulation experiments, it was demonstrated that the proposed methods have sure screening and ranking consistency properties. Finally, the results from simulation and real-dataset experiments show that in feature screening, the proposed methods investigated are robust in performance and faster in computational speed compared with an existing method.

DEVELOPMENT OF A HYBRID MODEL FOR THREE-DIMENSIONAL GIS

This paper presents a hybrid model for three_dimensional Geographical Information Systems which is an integration of surface_ and volume_based models.The Triangulated Irregular Network (TIN) and octree models are integrated in this hybrid model.The TIN model works as a surface_based model which mainly serves for surface presentation and visualization.On the other hand,the octree encoding supports volumetric analysis.The designed data structure brings a major advantage in the three_dimensional selective retrieval.This technique increases the efficiency of three_dimensional data operation.

A three-dimensional semi-implicit unstructured grid finite volume ocean model

A new three-dimensional semi-implicit finite-volume ocean model has been developed for simulating the coastal ocean circulation, which is based on the staggered C-unstructured non-orthogonal grid in the hor- izontal direction and z-level grid in the vertical direction. The three-dimensional model is discretized by the semi-implicit finite-volume method, in that the free-surface and the vertical diffusion are semi-implicit, thereby removing stability limitations associated with the surface gravity wave and vertical diffusion terms. The remaining terms in the momentum equations are discretized explicitly by an integral method. The partial cell method is used for resolving topography, which enables the model to better represent irregular topography. The model has been tested against analytical cases for wind and tidal oscillation circulation, and is applied to simulating the tidal flow in the Bohal Sea. The results are in good agreement both with the analytical solutions and measurement results.

Three-Dimension Mathematical Model of Total Phosphor in the Reservoir and Application

Taking the transport of total phosphor pollutants in the Beijing Miyun reservoir for example,we have obtained three dimensional distributing regularity of total phosphor pollutants by the calculation of the linear interpolation value of each point between horizontal layers. The credibility analysis in allusion to this method was carried out and the programming scheme for realizing this method was set forth.

Model-based Predictive Control for Spatially-distributed Systems Using Dimensional Reduction Models

In this paper, a low-dimensional multiple-input and multiple-output （MIMO） model predictive control （MPC） configuration is presented for partial differential equation （PDE） unknown spatially-distributed systems （SDSs）. First, the dimension reduction with principal component analysis （PCA） is used to transform the high-dimensional spatio-temporal data into a low-dimensional time domain. The MPC strategy is proposed based on the online correction low-dimensional models, where the state of the system at a previous time is used to correct the output of low-dimensional models. Sufficient conditions for closed-loop stability are presented and proven. Simulations demonstrate the accuracy and efficiency of the proposed methodologies.

Simulation of two-dimensional interior ballistics model of solid propellant electrothermal-chem ical launch with discharge rod plasma generator

Instead of the capillary plasma generator(CPG),a discharge rod plasma generator(DRPG)is used in the30 mm electrothermal-chemical(ETC)gun to improve the ignition uniformity of the solid propellant.An axisymmetric two-dimensional interior ballistics model of the solid propellant ETC gun(2D-IB-SPETCG)is presented to describe the process of the ETC launch.Both calculated pressure and projectile muzzle velocity accord well with the experimental results.The feasibility of the 2D-IB-SPETCG model is proved.Depending on the experimental data and initial parameters,detailed distribution of the ballistics parameters can be simulated.With the distribution of pressure and temperature of the gas phase and the propellant,the influence of plasma during the ignition process can be analyzed.Because of the radial flowing plasma,the propellant in the area of the DRPG is ignited within 0.01 ms,while all propellant in the chamber is ignited within 0.09 ms.The radial ignition delay time is much less than the axial delay time.During the ignition process,the radial pressure difference is less than 5 MPa at the place 0.025 m away from the breech.The radial ignition uniformity is proved.The temperature of the gas increases from several thousand K(conventional ignition)to several ten thousand K(plasma ignition).Compare the distribution of the density and temperature of the gas,we know that low density and high temperature gas appears near the exits of the DRPG,while high density and low temperature gas appears at the wall near the breech.The simulation of the 2D-IB-SPETCG model is an effective way to investigate the interior ballistics process of the ETC launch.The 2D-IB-SPETC model can be used for prediction and improvement of experiments.

Seismic data reconstruction based on low dimensional manifold model

Seismic data reconstruction is an essential and yet fundamental step in seismic data processing workflow,which is of profound significance to improve migration imaging quality,multiple suppression effect,and seismic inversion accuracy.Regularization methods play a central role in solving the underdetermined inverse problem of seismic data reconstruction.In this paper,a novel regularization approach is proposed,the low dimensional manifold model(LDMM),for reconstructing the missing seismic data.Our work relies on the fact that seismic patches always occupy a low dimensional manifold.Specifically,we exploit the dimension of the seismic patches manifold as a regularization term in the reconstruction problem,and reconstruct the missing seismic data by enforcing low dimensionality on this manifold.The crucial procedure of the proposed method is to solve the dimension of the patches manifold.Toward this,we adopt an efficient dimensionality calculation method based on low-rank approximation,which provides a reliable safeguard to enforce the constraints in the reconstruction process.Numerical experiments performed on synthetic and field seismic data demonstrate that,compared with the curvelet-based sparsity-promoting L1-norm minimization method and the multichannel singular spectrum analysis method,the proposed method obtains state-of-the-art reconstruction results.

High—Dimensional Integrable Models with Infinitely Dimensional Virasoro—Type Symmetry Algebra

Using every realization of the Virasoro-type symmetry algebra [o(f1) o(f2)] = σ(f1f2 - f2f1), we can obtain various high-dimensional integrable models under the meaning that they possess infinitely many symmetries. By means of a concrete realization, many (3 + 1)-dimensional equations which possess Kac Moody-Virasoro-type infinite dimensional symmetry algebras are obtained.

Modeling and validating three dimensional human normal cervix and cervical cancer tissues in vitro

The use of three dimensional in vitro systems in cancer research is a promising path for developing effective anticancer therapies.The aim of this study was to engineer a functional 3-D in vitro model of normal and cancerous cervical tissue.Normal epithelial and immortalized cervical epithelial carcinoma cell lines were used to construct 3-D artificial normal cervical and cervical cancerous tissues.De-epidermised dermis（DED） was used as a scaffold for both models.Morphological analyses were conducted by using hematoxylin and eosin staining and characteristics of the models were studied by analyzing the expression of different structural cytokeratins and differential protein marker MAX dimerisation protein 1（Mad1） using immunohistochemical technique.Haematoxylin and eosin staining results showed that normal cervical tissue had multi epithelial layers while cancerous cervical tissue showed dysplastic changes.Immunohistochemistry staining revealed that for normal cervix model cytokeratin 10 was expressed in the upper stratified layer of the epithelium while cytokeratin 5 was expressed mainly in the middle and basal layer.Cytokeratin 19 was weakly expressed in a few basal cells.Cervical cancer model showed cytokeratin 19 expression in different epithelial layers and weak or no expression for cytokeratin 5 and cytokeratin 10.Madl expression was detected in some suprabasal cells.The 3-D in vitro models showed stratified epithelial layers and expressed the same types and patterns of differentiation marker proteins as seen in corresponding in vivo tissue in either normal cervical or cervical cancerous tissue.These findings imply that they can serve as functional normal and cervical cancer models.

A 3-DIMENSIONAL DATA MODEL FOR VISUALIZING CLOVERLEAF JUNCTION IN A CITY MODEL

The research work has been seldom done about cloverleaf junction expression in a 3-dimensional city model (3DCM). The main reason is that the cloverleaf junction is often in a complex and enormous construction. Its main body is bestraddle in air,and has aerial intersections between its parts. This complex feature made cloverleaf junction quite different from buildings and terrain, therefore, it is difficult to express this kind of spatial objects in the same way as for buildings and terrain. In this paper,authors analyze spatial characteristics of cloverleaf junction, propose an all-constraint points TIN algorithm to partition cloverleaf junction road surface, and develop a method to visualize cloverleaf junction road surface using TIN. In order to manage cloverleaf junction data efficiently, the authors also analyzed the mechanism of 3DCM data management, extended BLOB type in relational database, and combined R-tree index to manage 3D spatial data. Based on this extension, an appropriate data

Risk Assessment of Coal Mine Water Hazard Based on Three-Dimensional Model of Geology and Underground Space

Mine safety have top-five disasters,which including the water,gas,fire,dust and geological dynamic disaster.The coal mine water disaster is one of the important factors which restricted the development of China’s coal production.It is showed by statistics that 60%of mine accidents are affected by groundwater,which not only result in the production losses,casualties and a variety of

A NEW UNSTEADY THREE DIMENSIONAL MODEL FOR MACROMOLECULAR TRANSPORT AND WATER FILTRATION ACROSS THE ARTERIAL WALL

A new unsteady three-dimensional convective-diffusive mathematical model for the transportation of macromolecules and water across the arterial wall was proposed . After the formation of leaky junctions due to the mitosis of endothelial cell of the arterial wall, the macromolecular transport happens surrounding the leaky cells. The arterial wall was divided into four layers: the endothelial layer, the subendothelial intima, the internal elastic lamina and the media for the convenience of research. The time-dependent concentration growth,the effect of the shape of endothelial cell and the effect of physiological parameters were analyzed. The analytical solution of velocity field and pressure field of water flow across the arterial wall were obtained; and concentration distribution of three macromolecules ; LDL,HRP and Albumin, were calculated with numerical simulation method. The new theory predicts, the maximum and distribution areas of time dependent concentration with round shape endothelial cell are both larger than that with ellipse-shape endothelial cell. The model also predicts the concentration growth is much alike that of a two-dimensional model and it shows that the concentration reaches its peak at the leaky junction where atherosclerotic formation frequently occurs and falls down rapidly in a limited area beginning from its earlier time growth to the state when macromolecular transfer approaches steadily. These predictions of the new model are in agreement with the experimental observation for the growth and concentration distribution of LDL and Albumin.

The study on three-dimensional numerical model and fronts of the Jiulong Estuary and the Xiamen Bay

Applying the methods of on-site observation and dynamic model, the research on the fronts at the Jiulong Estuary has been carried out, during which spatial and temporal distribution, dynamic characteristics and formation mechanism of salinity fronts are analyzed and discussed. The research shows that the estuarine fronts mainly lie in the area from the Jiyu Islet to the Haimen Island, outside of Yuweizai to Hulishan cross-section, the near coast of Yuweizai and the south of the Songyu-Gulangyu Channel. The fronts in the former two regions are formed directly by plume, while the one near the coast of Yuweizai is a tidal intrusion front caused by flood current and the one at the south of the Songyu-Gulangyu Channel is the result of current shear transformation. Under normal circumstances, fresh water of the Jiulong River mainly influences the inside of the Xiamen Bay, and when it is in typhoon seasons, plume front can affect the Taiwan Strait and has an effect on the biogeochemical Drocesses in the strait.

Simulation of quasi-dimensional model for diesel engine working process

In order to satisfy the demand of validity and real time operating performance of diesel engine model used in hardware-in-the-loop simulation system,a simplified quasi-dimensional model for diesel engine working process was proposed,which was based on the phase-divided spray mixing model.The software MATLAB/Simulink was utilized to simulate diesel engine performance parameters.The comparisons between calculated results and experimental data show that the relative error of power and brake specific fuel consumption is less than 2.8%,and the relative error of nitric oxide and soot emissions is less than 9.1%.At the same time,the average computational time for simulation of one working process with the new model is 36 s,which presents good real time operating performance of the model.The simulation results also indicate that the nozzle flow coefficient has great influence on the prediction precision of performance parameters in diesel engine simulation model.

A Three-dimensional Heterogeneous Mass Transfer Model for the Absorption of Gas into Multiphase System

一个三维的异构的传质模型被建议在气体吸收上调查驱散的粒子的改进。计算本地、全面的改进因素的策略被建议。而不是一个全球格子，合成重叠格子被采用，它简化三维的模型方程的安装和解决方案。驱散的粒子停顿，粒子大小，搬运部件的液体固体分配系数，典型接触时间，和在粒子 andgas 液体之间的最短的距离连接，这被发现在吸收上有主要影响改进因素。气体吸收和在液膜的搬运部件的径向扩散上的 particle-particleinteraction 我们与多粒子模拟再学习。建议模型相当从文学预言了试验数据很好。

Three-Dimensional Model Test for Port Engineering with Multi-Directional Waves

-The necessity of using irregular waves, especially multi- directional waves to conduct three-dimensional model tests for port engineering and the test method are described in this paper through an example of model test for a port. The test results show that a deep navigation channel has a large effect on the waves in front of the breakwater near the port entrance and on the wave condition in the port.

Numerical Simulation for Two-Phase Water Hammer Flows in Pipe by Quasi-Two-Dimensional Model

The features of a quasi-two-dimensional( quasi-2D) model for simulating two-phase water hammer flows with vaporous cavity in a pipe are investigated. The quasi-2D model with discrete vaporous cavity in the pipe is proposed in this paper. This model uses the quasi-2D model for pure liquid zone and one-dimensional( 1D) discrete vapor cavity model for vaporous cavity zone. The quasi-2D model solves two-dimensional equations for both axial and radial velocities and 1D equations for both pressure head and discharge by the method of characteristics. The 1D discrete vapor cavity model is used to simulate the vaporous cavity occurred when the pressure in the local pipe is lower than the vapor pressure of the liquid. The proposed model is used to simulate two-phase water flows caused by the rapid downstream valve closure in a reservoir-pipe-valve system.The results obtained by the proposed model are compared with those by the corresponding 1D model and the experimental ones provided by the literature,respectively. The comparison shows that the maximum pressure heads simulated by the proposed model are more accurate than those by the corresponding 1D model.

Performance assessment of two-dimensional hydraulic models for generation of flood inundation maps in mountain river basins

Hydraulic models for the generation of flood inundation maps are not commonly applied in mountain river basins because of the difficulty in modeling the hydraulic behavior and the complex topography. This paper presents a comparative analysis of the performance of four twodimensional hydraulic models (HEC-RAS 2D, Iber 2D, Flood Modeller 2D, and PCSWMM 2D) with respect to the generation of flood inundation maps. The study area covers a 5-km reach of the Santa B-arbara River located in the Ecuadorian Andes, at 2330 masl, in Gualaceo. The model's performance was evaluated based on the water surface elevation and flood extent, in terms of the mean absolute difference and measure of fit. The analysis revealed that, for a given case, Iber 2D has the best performance in simulating the water level and inundation for flood events with 20- and 50-year return periods, respectively, followed by Flood Modeller 2D, HEC-RAS 2D, and PCSWMM 2D in terms of their performance. Grid resolution, the way in which hydraulic structures are mimicked, the model code, and the default value of the parameters are considered the main sources of prediction uncertainty.