Logistic回归与分类树模型的比较

为了有效地评估客户的可信度,提高信贷机构经济效益。文中通过统计学中的参数方法Logistic回归和非参数方法分类树这两种方法,建立两种模型对数据进行预测,应用SPSS软件的Binary Logistic Regression方法,利用ROC曲线的性质来对模型的性能进行评价,根据输出结果比较两种模型在应用中都是可行有效的,在实际操作中应因地制宜,把握两种方法的优势,得到更有价值的结果.     

X射线在炸药柱中衰减规律的试验研究

为了对武器进行X射线无损检测，必需掌握X射线在炸药中的衰减规律。通过试验测定，利用线性回归方法对不同情况下测得的X射线在某炸药中的穿透率进行拟合，分析拟合结果得出X射线在某炸药中的衰减规律。研究结果表明，X射线在某炸药中的衰减是比较明显的，其衰减规律符合三次方程形式。     

An efficient method for solving the Signorini problem in the simulation of free-form surfaces produced by belt grinding

Industrial robots are recently introduced to the belt grinding of free-form surfaces to obtain high productive efficiency and constant surface quality. The simulation of belt grinding process can facilitate planning grinding paths and writing robotic programs before manufacturing. In simulation, it is crucial to get the force distribution in the contact area between the workpiece and the elastic contact wheel because the uneven distributed local forces are the main reason to the unequal local removals on the grated surface. The traditional way is to simplify this contact problem as a Signorini contact problem and use the finite element method (FEM) to calculate the force distribution. However, the FEM model is too computationally expensive to meet the real-time requirement. A new model based on support vector regression (SVR) technique is developed in this paper to calculate the force distribution instead of the FEM model. The new model approximates the FEM model with an error smaller than 5%, but executes much faster (1 s vs 15 min by FEM). With this new model, the real-time simulation and even the on-line robot control of grinding processes can be further conducted.     

多软件平台联合下的箱体结构优化设计

本文通过对某型号压缩机箱体的研究，在多种软件的辅助分析下，通过采用最小二乘逼进理论及MATLAB这一数学工具，建立了以影响箱体重量的五个几何尺寸为设计变量、以箱体重量为目标函数、以箱体承受应力为约束的优化模型。通过优化设计软件，求得了其最优解。在求解的过程中，通过多种途径对建模、优化进行了验证。     

Development of Habitat Suitability Criteria and In‐Stream Habitat Assessment for the Benthic Cyanobacteria Phormidium

Global demand for freshwater has led to unprecedented levels of water abstraction from riverine systems. This has resulted in large alterations in natural river flows. The deleterious impacts of reduced flows on fish and macroinvertebrate abundances have been thoroughly investigated; in contrast, there is a limited understanding of the potential for changes in the abundance of nuisance benthic algal/cyanobacterial blooms. In New Zealand, Phormidium sp. blooms are common in numerous rivers during summer low flows. In this study, an in‐stream habitat assessment is used to examine the relationship between Phormidium habitat availability and reducing flows. Over 650 observations of Phormidium mats, from seven sites (Hutt River, lower North Island, New Zealand), were used to construct habitat suitability curves for depth, velocity and substrate. Preference curves were fitted using both the ‘forage ratio’ and ‘quantile regression’ methods. Phormidium growth, observed at all seven sites, increased significantly from upstream (uppermost site, 5.2% mat cover) to downstream (63.5%). The habitat suitability curves revealed Phormidium had a large tolerance to velocity, depth and substrate type. Consequently, decreases in flow had only negligible effects on available Phormidium habitat. During periods of stable flow, Phormidium abundance positively correlated with increased nitrogen concentrations, potentially explaining the large variation in Phormidium cover from upstream to downstream. Quantile regression generated habitat suitability criteria were a more accurate predictor of available Phormidium habitat than the forage ratio criteria. Copyright © 2013 John Wiley & Sons, Ltd.     

Ongoing verification of a multiphysics community code: FLASH

When developing a complex, multi‐authored code, daily testing on multiple platforms and under a variety of conditions is essential. It is therefore necessary to have a regression test suite that is easily administered and configured, as well as a way to easily view and interpret the test suite results. We describe the methodology for verification of FLASH, a highly capable multiphysics scientific application code with a wide user base. The methodology uses a combination of unit and regression tests and an in‐house testing software that is optimized for operation under limited resources. Although our practical implementations do not always comply with theoretical regression‐testing research, our methodology provides a comprehensive verification of a large scientific code under resource constraints.Copyright © 2013 John Wiley & Sons, Ltd.     

基于逻辑回归模型的木片和树皮的图像识别

为了提高木片生物能的利用率,需要除去木片中的树皮杂质。识别木片和树皮的图像对于实现二者分离具有重要作用。提出一种基于逻辑回归模型来对木片和树皮进行图像识别的算法。算法流程包括计算各种可以描述木片和树皮纹理的参数,利用SAS统计分析出最佳参数,将最佳参数计算所得的样本数据用于SAS LOGISTIC过程得到预测方程,再将预测方程用于逻辑回归模型来实现木片和树皮的图像识别。通过实验可知,最佳参数为灰度方差,该图像识别算法的正确率能达到97.4%。     

基于传感器的飞机地面结霜实验分析与预测

飞机在地面的结霜会破坏飞机的气动外形,对飞机的飞行性能产生不利的影响。为了对地面飞机结霜进行准确监测与预测,利用多种传感器对机翼的结霜环境进行实验研究,通过分析不同环境下模拟机翼的霜层生长曲线,探寻和统计飞机机翼的地面结霜规律。对结霜实验的数据进行分析,提出了不同结霜环境下的多项式回归预测模型。同时将多项式进行正交化,避免了多重共线性,使预测模型的曲线拟合程度高,预测值具有较高的精度,实现了对飞机的地面结霜进行快速精确的预测。     

高分辨TerraSAR-X 图像舰船目标几何参数提取方法

准确提取SAR图像舰船目标几何参数是实现舰船分类的基础。由于SAR特有的成像机理, SAR图像旁瓣效应和方位模糊严重影响了舰船目标的几何参数估计,为此提出一种基于迭代线性回归的舰船几何参数提取方法。首先通过线性回归提取目标的主轴方位;然后利用循环迭代剔除远离主轴的虚假目标像素;最后用实测SAR数据进行了验证。实验结果表明,该算法对高分辨率的SAR舰船目标切片具有较好的参数估计效果。