基于超图稀疏的低秩属性选择算法用于多回归分析

针对已有的多回归算法未能更好地考虑不同列类别之间的关联性,提出了一种基于超图稀疏的属性选择方法。利用稀疏理论中的l_(2,p)-范数来改进线性回归模型,结合低秩结构和超图表示来分别考虑同一类别间的局部结构和不同类别间的全局结构,通过子空间学习中的线性判别分析(LDA)来进一步调整重构系数矩阵的结果。经实验验证,提出的算法在六个公开数据集上相比四种对比算法,在多回归分析中均能取得更好的结果。

某电站地下厂房区域初始地应力反演分析

根据某电站地下厂房区域工程地质条件以及实测地应力资料,分析影响区域初始地应力的因素,建立电站厂房区域三维有限元模型,利用FLAC^(3D)数值计算软件对工程区域初始地应力场进行三维有限元模拟,在实测地应力值与计算地应力值之间建立多元回归模型,求出各回归系数,结果表明模拟得到的初始地应力场与实测地应力值相符,能够满足工程设计、施工的需要.

糖尿病肾病患者血小板GMP-140水平变化及其相关因素的研究

探讨糖尿病患者血小板GMP - 140含量变化及其相关因素与糖尿病肾病发生发展的关系。本文选取糖尿病无肾脏病组 (DM组 ) 144例 (平均病程 2 5 .5± 18.6个月 ) ,糖尿病有肾脏病变组 (DN组 ) 80例 (平均病程5 8.7± 31.6个月 ) ,以及 5 0名对照组。用放免法分别测定其血小板GMP - 140和血α1-MG、β2 -MG ;层析柱法测HBA1C ;生化法测FBG、PBG、Ch、TG、HDL、FG ;放免法测 2 4h尿ALB、IgG、α1-MG和 β2 -MG。经计算机软件处理测定结果 ,进行t检验 ,并对结果进行直线相关和多因素分析。结果表明DN组血小板GMP - 140、FG、DBP、TG、HBA1C和PBG水平 ,明显高于DM组和对照组 (P <0 .0 5 - 0 .0 0 1) ;DN组的HDL、SBP和FBG与DM组无显著性差异 (P均 >0 .0 5 ) ,高于对照组 (P <0 .0 1- 0 .0 0 1)。对 80例DN患者的血小板GMP - 140与TG、DBP和HBA1C等 ,作多因素回归分析 (P <0 .0 1)。本文提示 :(1)血小板GMP - 140含量明显增高 ,能够准确反映血小板活化、活化程度及血栓形成倾向 ,以及血小板大量活化 ,参与了肾脏微血管病变。 (2 )多因素回归分析发现 ,血TG、Ch和HBA1C是引起血小板活性增强的独立因素 ,Ⅱ型糖尿病患者的脂质代谢紊乱使血小板活性增强。 (3)血压升高可加速DN的形成和发展 ,对血小板GMP - 140含量?

Prediction of pre-oxidation efficiency of refractory gold concentrate by ozone in ferric sulfate solution using artificial neural networks

An artificial neural network model was developed to predict the oxidation of refractory gold concentrate （RGC） by ozone and ferric ions. The concentration of ozone and ferric ions, pulp density, oxygen amount, leaching time and temperature were employed as inputs to the network; the output of the network was the percentage of the ferric extraction iron from RGC. The multilayered feed-forward networks were trained by 33 sets of input-output patterns using a back propagation algorithm; a three-layer network with 8 neurons in the hidden layer gave optimal results. The model gave good predictions of high correlation coefficient （R2=0.966）. The predictions by ANN are more accurate when compared with conventional multivariate regression analysis （MVRA）. In addition, calculation with ANN model indicates that temperature is the predominant parameter and ozone concentration is the lesser influential parameter in the pre-oxidation process of refractory gold ore. The ANN neural network model accurately estimates the ferric extraction during pretreatment process of RGC in gold smelter plants and can be used to optimize the process parameters.

Extraction of Soil Organic Matter Information by Multi-spectral Remote Sensing Based on Diverse Landforms

Based on diverse landforms, the correlation between soil organic matter content and multi-spectral band of remote sensing image was analyzed in this pa- per. In addition, the inversion models were built for the soil organic matter content in different landforms. The results showed that the spectral reflectance was nega- tively related to soil organic matter content; linear regression analysis of remove was performed throughout the bands using SPSS. When the inversion models were built based on all the bands, better fitting effect was obtained. The precision of in- version models built based on different landforms was higher than those built re- gardless landforms. Compared with the actual value, the identification level of soil organic matter content was 91 65% under the allowable error was 7%. It indicated that the extraction of soil organic matter with inversion model that was built based on different landforrrs was feasible with higher precision.

Analyses of unified congestion measures for interrupted traffic flow on urban roads

To study the congestion of interrupted flow on urban roads, a comprehensive evaluation method is proposed. First, based on the results of correlation analysis between different parameters of interrupted flow, the traffic parameters of interrupted traffic flow are divided into two categories： the basic parameters and the operation parameters. Polynomial regression is used to formulize the nonlinear relationships between the basic parameters and the operation parameters. Then, the congestion model incorporating both operational and volume characteristics of traffic flow is proposed. The inputs of the model are the basic parameters, while the output is a dimensionless index value between 0 and 1. Finally, the proposed methods are compared with existing evaluation measures of congestion. Results show that the proposed indices can capture the variation of both the basic parameters and the operation parameters, which is more balanced compared with the existing evaluation measures.

New Math Model for Designing Engine Radiator

Aim To raise the H-NE math model for designing engine radiators.Methods Thismodel was founded on multi-nonlinear regression technology,and the quadratic equatingmethod was used to smooth original data.Through the sensitivity analysis about radiatoroptimization designing,some major restraints were obtained.Results Type 1301C radiator wasoptimizingly renwed by use o this model,and the new radiator is more excellent than theold one in respect of the heat transfer factor and the manufacturing cost.Conclusion Theradiation performance can be improved,the area can be reduced,and the fitting scope can bewidened if this model is used in designing radiators.

Multivariate Regression Analysis on Correlated Characters about Fresh Pod Yield of Fresh Edible Soybean

Objective] The alm was to survey 10 characters of 8 fresh edibIe soy-bean varieties, analyze maln Ioading factors using principal component analysis, and estabIish muItipIe regression equation on fresh pod yield. [Methods] Through princi-pal component analysis on 10 characters of 8 fresh edibIe soybean varieties, char-acters reIated to fresh pod yield of fresh edibIe soybean were cIarified. [Results] Af-ter the principal components analysis, pod weight per pIant, 100-seed weight and pod number per pIant of fresh edibIe soybean were chosen to study their reIation with the yield of fresh edibIe soybean, moreover, it was demonstrated that the reIa-tion was Iinear reIation, thus it was suitabIe for muItivariate regression analysis. Fi-nal y, the mathematical expression formuIa about fresh pod yield was estabIished. [Conclusions] There were three characters affecting fresh pod yield, nameIy, pod weight per pIant, 100-seed weight and pod number per pIant, the mathematical equation was y=816.732＋4.145X6-0.718X8-0.985X9 （X6： pod weight per pIant; X8： 100-seed weight; X9： pod number per pIant）.

Associations of big endothelin-1 and C-reactive protein in atrial fibrillation

Background Atrial fibrillation （AF） is associated with inflammation and endothelial dysfunction. However, the association between inflammation （as indexed by high-sensitivity C-reactive protein, hs-CRP） and endothelial function [as indexed by big endothelin-1 （ET-1）] in AF patients remains unclear. Methods We enrolled 128 patients with lone AF, among which 83 had paroxysmal AF, and 45 had persistent AF. Eighty-two age- and gender-matched controls of paroxysmal supraventricular tachycardia without AF history were evaluated. Plasma hs-CRP, big ET-1 levels and other clinical characteristics were compared among the groups. Results Patients with persistent AF had higher hs-CRP concentrations than those with paroxysmal AF （P 〈 0.05）, both groups had higher hs-CRP level than controls （P 〈 0.05）. Patients with persistent AF had higher big ET-1 level than those with paroxysmal AF, although the difference did not reach the statistical significance （P 〉 0.05）, and both groups had higher big ET-1 levels than controls （P 〈 0.05）. Multiple regression analyses revealed hs-CRP as an inde- pendent determinant of AF （P 〈 0.001）. Further adjusted for big ET-1, both big ET-1 and hs-CRP were independent predictors for AF （P 〈 0.001）, but the odds ratio for hs-CRP in predicting AF attenuated from 8.043 to 3.241. There was a positive relation between hs-CRP level and big ET-1 level in paroxysmal AF patients （r = 0.563, P 〈 0.05）, however, the relationship in persistent AF patients was poor （r = 0.094, P 〈 0.05）. Conclusions Both plasma hs-CRP and big ET-1 levels are elevated in lone AF patients, and are associated with AF. In paroxysmal lone AF patients, there were significant positive correlations between plasma hs-CRP level and big ET- 1 level.

Differential effects of physical activity and sleep duration on cognitive function in young adults

Purpose: Although exercise and sleep duration habits are associated with cognitive function, their beneficial effects on cognitive function remain unclear. We aimed to examine the effect of sleep duration and daily physical activity on cognitive function, elucidating the neural mechanisms using near-infrared spectroscopy(NIRS).Methods: A total of 23 healthy young adults(age 22.0 ± 2.2 years) participated in this study. Exercise amount was assessed using a uniaxial accelerometer. We evaluated total sleep time(TST) and sleep efficiency by actigraphy. Cognitive function was tested using the N-back task, the Wisconsin Card Sorting Test(WCST), and the Continuous Performance Test—Identical Pairs(CPT-IP), and the cortical oxygenated hemoglobin levels during a word fluency task were measured with NIRS.Results: Exercise amount was significantly correlated with reaction time on 0- and 1-back tasks(r = —0.602, p = 0.002; r = —0.446, p = 0.033,respectively), whereas TST was significantly correlated with % corrects on the 2-back task(r = 0.486, p = 0.019). Multiple regression analysis,including exercise amount, TST, and sleep efficiency, revealed that exercise amount was the most significant factor for reaction time on 0- and 1-back tasks(b = —0.634, p = 0.002; b = —0.454, p = 0.031, respectively), and TST was the most significant factor for % corrects on the 2-back task(b = 0.542, p = 0.014). The parameter measured by WCST and CPT-IP was not significantly correlated with TST or exercise amount. Exercise amount, but not TST, was significantly correlated with the mean area under the NIRS curve in the prefrontal area(r = 0.492, p = 0.017).Conclusion: Exercise amount and TST had differential effects on working memory and cortical activation in the prefrontal area. Daily physical activity and appropriate sleep duration may play an important role in working memory.

Relationship between red blood cell distribution width and intermediate-term mortality in elderly patients after percutaneous coronary intervention

Background Large-scale clinical research on the relationship between red blood cell distribution width （RDW） and intermediate-term prognosis in elderly patients with coronary artery disease （CAD） is lacking. Thus, this study investigated the effects of RDW on the intermediate-term mortality of elderly patients who underwent elective percutaneous coronary intervention （PCI）. Methods Data from 1891 patients 〉 65 years old underwent elective PCI from July 2009 to September 2011 were collected. Based on preoperative median RDW （12.3%）, the patients were divided into two groups. The low RDW group （RDW 〈 12.3%） had 899 cases; the high RDW group （RDW 〉 12.3%） had 992 cases. The all-cause mortality rates of the two groups were compared. Results Patients in the high RDW group were more likely to be female and accompanied with diabetes, had lower hemoglobin level. The mean follow-up period was 527 days. During follow-up, 61 patients died （3.2%）. The postoperative mortality of the high RDW group was significantly higher than that of the low RDW group （4.3% vs. 2.0%, P = 0.004）. After adjusting other factors, multivariate Cox regression analysis revealed that preoperative high RDW was significantly associated with postoperative all-cause mortality （hazard ratio： 2.301, 95% confidence interval： 1.106-4.785, P = 0.026）. Conclusions Increased RDW was an independent predictor of the increased intermediate-term all-cause mortality in elderly CAD patients after elective PCI.

Changing Rural Development Inequality in Jilin Province, Northeast China

Rural development inequality is an important practical issue during the course of full establishment of a ′moderately well-off society′ in modern China,and the objective understanding and evaluation of the status and regional inequality of rural development can provide scientific basis for ′building a new countryside′ and coordination development of rural-urban regions.Based on the county-level data of 2000,2005 and 2009,this paper examines the rural development inequality of Jilin Province in Northeast China by establishing a rural development index.The spatio-temporal dynamic patterns and domain factors are discussed by using the method of exploratory spatial data analysis and multi-regression model.The results are shown as follows.Firstly,most of the counties were in lower development level,which accounted for 58.3%,62.5% and 66.7% of the total counties in 2000,2005 and 2009,respectively.The characteristics of spatial inequality were very obvious at county level.For example,rural development level of Changchun Proper and the proper of seven prefecture-level cities were much higher than that of the surrounding regions.The counties in the eastern and northern Jilin Province were the lowest regions of rural development level,while the middle counties were the rapid growth areas in rural economy.Secondly,Moran′s I of rural development index(RDI) was 0.01,–0.16 and –0.06 in 2000,2005 and 2009,respectively,which indicated that spatial agglomeration of RDI was not obvious in Jilin Province,and took on the characteristic of random distribution.The counties of both the units and its adjacent units have higher development level(HH) were transferred from the western areas to the eastern areas,while the countries of both the units and its adjacent units have lower development level(LL) were diffused from the eastern to middle and western Jilin Province.Finally,the result of multi-regression analysis showed that the improvement of agricultural production condition,development of agricultural economics and the adjustment of industrial structure were the domain factors affecting rural development inequality of Jilin Province in the later ten years.

Correlations between endoscopic and clinical disease activity indices in intestinal Behcet's disease

AIM:To develop a novel endoscopic severity model of intestinal Behcet's disease(BD) and to evaluate its feasibility by comparing it with the actual disease activity index for intestinal Behcet's disease(DAIBD).METHODS:We reviewed the medical records of 167 intestinal BD patients between March 1986 and April 2011.We also investigated the endoscopic parameters including ulcer locations,distribution,number,depth,shape,size and margin to identify independent factors associated with DAIBD.An endoscopic severity model was developed using significant colonoscopic variables identified by multivariate regression analysis and its correlation with the DAIBD was evaluated.To determine factors related to the discrepancy between endoscopic severity and clinical activity,clinical characteristics and laboratory markers of the patients were analyzed.RESULTS:A multivariate regression analysis revealed that the number of intestinal ulcers(≥ 2,P = 0.031) and volcanoshaped ulcers(P = 0.001) were predictive factors for the DAIBD.An endoscopic severity model(Y) was developed based on selected endoscopic variables as follows:Y = 47.44 + 9.04 × non-Ileocecal area + 11.85 ×≥ 2 of intestinal ulcers + 5.03 × shallow ulcers + 12.76 × deep ulcers + 4.47 × geographicshaped ulcers + 26.93 × volcano-shaped ulcers + 8.65 ×≥ 20 mm of intestinal ulcers.However,endoscopic parameters used in the multivariate analysis explained only 18.9% of the DAIBD variance.Patients with severe DAIBD scores but with moderately predicted disease activity by the endoscopic severity model had more symptoms of irritable bowel syndrome(21.4% vs 4.9%,P = 0.026) and a lower rate of corticosteroid use(50.0% vs 75.6%,P = 0.016) than those with severe DAIBD scores and accurately predicted disease by the model.CONCLUSION:Our study showed that the number of intestinal ulcers and volcano-shaped ulcers were predictive factors for severe DAIBD scores.However,the correlation between endoscopic severity and DAIBD(r = 0.434) was weak.

A method for real power transfer allocation using multivariable regression analysis

A multivariable regression(MVR) approach is proposed to identify the real power transfer between generators and loads.Based on solved load flow results,it first uses modified nodal equation method(MNE) to determine real power contribution from each generator to loads.Then,the results of MNE method and load flow information are utilized to determine suitable regression coefficients using MVR model to estimate the power transfer.The 25-bus equivalent system of south Malaysia is utilized as a test system to illustrate the effectiveness of the MVR output compared to that of the MNE method.The error of the estimate of MVR method ranges from 0.001 4 to 0.007 9.Furthermore,when compared to MNE method,MVR method computes generator contribution to loads within 26.40 ms whereas the MNE method takes 360 ms for the calculation of same real power transfer allocation.Therefore,MVR method is more suitable for real time power transfer allocation.

Soil Wettability as Affected by Soil Characteristics and Land Use

Depth distribution of soil wettability and its correlations with vegetation type, soil texture, and pH were investigated under various land use (cropland, grassland, and forestland) and soil management systems. Wettability was evaluated by contact angle with the Wilhelmy plate method. Water repellency was likely to be present under permanently vegetated land, but less common on tilled agricultural land. It was mostly prevalent in the topsoil, especially in coarse-textured soils, and decreased in the subsoil. However, the depth dependency of wettability could not be derived from the investigated wide range of soils. The correlation and multiple regression analysis revealed that the wettability in repellent soils was affected more by soil organic carbon (SOC) than by soil texture and pH, whereas in wettable soils, soil texture and pH were more effective than SOC. Furthermore, the quality of SOC seemed to be more important in determining wettability than its quantity, as proofed by stronger hydrophobicity under coniferous than under deciduous forestland. Soil management had a minor effect on wettability if conventional and conservation tillage or different grazing intensities were considered.

Stability of mine ventilation system based on multiple regression analysis

In order to overcome the disadvantages of diagonal connection structures that are complex and for which it is difficult to derive the discriminant of the airflow directions of airways, we have applied a multiple regression method to analyze the effect, of changing the rules of mine airflows, on the stability of a mine ventilation system. The amount of air （ Qj ） is determined for the major airway and an optimum regression equation was derived for Qi as a function of the independent variable （ Ri ）, i.e., the venti- lation resistance between different airways. Therefore, corresponding countermeasures are proposed according to the changes in airflows. The calculated results agree very well with our practical situation, indicating that multiple regression analysis is simple, quick and practical and is therefore an effective method to analyze the stability of mine ventilation systems.

Clinical outcomes in patients with ICU-related pancreatitis

AIM:To identify risk factors predictive of intensive care unit(ICU) mortality in patients with ventilator-related pancreatitis.The clinical outcomes of patients with ventilator-related pancreatitis were compared with those of patients with pancreatitis-related respiratory failure as well as controls.METHODS:One hundred and forty-eight patients with respiratory failure requiring mechanical ventilation and concomitant acute pancreatitis were identified from a prospectively collected dataset of 9108 consecutive patients admitted with respiratory failure over a period of five years.Sixty patients met the criteria for ventilator-related pancreatitis,and 88(control patients),for pancreatitis-related respiratory failure.RESULTS:Mortality rate in ventilator-related pancreatitis was comparable to that in ICU patients without pancreatitis by case-control methodology(P=0.544).Multivariate logistic regression analysis identified low PaO2/FiO2(OR:1.032,95% CI:1.006-1.059,P=0.016) as an independent risk factor for mortality in patients with ventilator-related pancreatitis.The mortality rate in patients with ventilator-related pancreatitis was lower than that in patients with acute pancreatitis-related respiratory failure(P<0.001).CONCLUSION:We found that low PaO2/FiO2 was an independent clinical parameter predictive of ICU mortality in patients with ventilator-related pancreatitis.

Detection of D2-40 monoclonal antibody-labeled lymphatic vessel invasion in esophageal squamous cell carcinoma and its clinicopathologic significance

Objective： This study aims to investigate the clinicopathologic significance of lymphatic vessel invasion （LVI） labeled by D2-40 monoclonal antibody in esophageal squamous cell carcinoma （ESCC）. Methods： Immunohistochemical assay was used to detect the expression of D2-40 and LVI in 107 ESCC patients. Then, the correlation between the clinicopathologic feature and the overall survival time of the patients was analyzed. Results： The lymph node metastasis rates were 70% and 21% in the LVI-positive and LVI-negative groups, respectively. The nodal metastasis rate was higher in the LVI-positive group than in the LVI-negative group. Multivariate regression analysis showed that LVI was related to nodal metastasis （P〈0.001）. The median survival time of the patients was 26 and 43 months in the LVI-positive and LVI-negative groups, respectively. Mthough univariate regression analysis showed significant difference between the two groups （P=0.014）, multivariate regression analysis revealed that LVI was not an independent prognostic factor for overall survival in the ESCC patients （P=0.062）. Lymphatic node metastasis （P=0.031）, clinical stage （P=0.019）, and residual tumor （P=0.026） were the independent prognostic factors. Conclusion： LVI labeled by D2-40 monoclonal antibody is a risk factor predictive of lymph node metastasis in ESCC patients.

Adherence to pharmacological and non-pharmacological treatment of frail hypertensive patients

Objective To investigate the relationship between frailty syndrome （FS） and adherence to pharmacological and non-pharmacological treatment for hypertension. Methods The study included 100 patients diagnosed with hypertension and treated with one or more hypotensive drugs. Results Frail patients obtained low scores （4.1 ± 2.0） for adherence to pharmaceutical treatment of hypertension, while non-frail patients obtained moderate scores （6.1 ± 2.1）. Non-frail patients had higher scores in two out of four domains of the Health Behavior Inventory （HBI）： positive mental attitudes （3.6 ± 0.4 vs. 3.2 ± 0.5; P = 0.006） and health practices （3.6 ± 0.5 vs. 3.2 ± 0.5; P 〈 0.03）; as well as higher global scores （HBI raw score）： 83.3 ±10.6 vs. 77.3 ± 9.5; P 〈 0.03. Multiple regression analysis showed that frailty syndrome （FS） was a statistically significant independent determinant of worse adherence to pharmacological treatment （β= -0.27; P 〈 0.001） and health behaviors （β = -0.10; P = 0.036）. Education was a statistically significant independent determinant of better adherence to pharmacological treatment （β = 0.82; P = 0.012）, while net income positively affected health behaviors as measured by the HBI （β = 0.39; P = 0.046）. Conclusions FS is a significant independent factor contributing to worse adherence to pharmacological and non-pharmacological treatment of hypertension. Better education significantly improves patients＇ adherence to the prescribed pharmacological treatment, while a good financial standing evidenced by high net income is a determinant of better adherence to health-related behaviors recommended in hypertension treatment.

Theoretical analysis of thermoviscoelastic contact between friction lining and wire rope in mine friction hoists

Serious accidents of mine hoists caused by high-speed sliding between friction lining and wire rope are often seen in coal mines.In order to solve this problem,we analyzed the contact characteristics between friction lining and wire rope.Then we carried out a dynamic mechanical analysis(DMA) to explain the change in mechanical properties of the friction lining as function of temperature and load frequency and found that temperature has a stronger effect on the mechanical properties than the frequency.We used multiple regression analysis to obtain the thermoviscoelastic constitutive relations of the friction lining.As well we derived the analytic solution for the thermoviscoelastic contact radius and pressure by combining the theory of viscoelastic contact mechanics with thermoviscoelastic constitutive relations.