Logistic回归在判别分析中的新用法

在非线性回归模型中,当响应变量为分类变量尤其是二分类变量时所采用的Logistic回归模型是一个很有应用前景的模型。它可以克服线性回归的一些缺点,提高判别分析的效率。通过分析Logistic回归应用于判别分析的特点提出采用二次判别法来提高判别效率,最后通过实际案例验证了所提做法的有效性。     

财务预警模型的变迁

根据国内外在财务预警模型研究中所运用的单变量判别法、多元线性回归分析、主成分分析、二元Logistic回归分析和神经网络分析等主要建模方法、应用情况和局限性，提出了财务预警模型研究中值得注意的若干问题和今后的发展趋势。     

大学毕业生职业心理适应预测模型研究

探讨评估和预测大学毕业生职业心理适应的技术与方法.使用霍兰德职业兴趣问卷、职业效能问卷等职业心理测验对725名大学毕业生进行调查,采用多元判别分析和人工神经网络建立大学毕业生职业心理适应的评估与预测模型.判别分析模型鉴别适应不良的毕业生回判正确率与交叉样本验证的正确率均达84％,对两组的判别正确率则达回判的79％和交叉验证的75％.利用BP人工神经网络对数据进行回代拟合,判别的正确率为93.2％;对验证集数据预测准确率为88.72％.与判别模型相比,神经网络模型的预测正确率更高.大学毕业生职业心理适应预测模型有助于快捷诊断和及早发现适应不良的毕业生,从而为大学生就业辅导工作提供重要参考与指导.     

盈余管理能否增加财务危机预警的预测能力——来自中国ST上市公司的证据

论文采用2003--2006年来自中国ST上市公司的数据,对于盈余管理是否增加财务危机预警的预测能力进行分析,即在财务信息、公司治理变量的预测模型中加入盈余管理变量,建立Logistic回归分析模型。研究发现：（1）财务危机公司相对正常公司,于财务危机发生前盈余管理程度较高。（2）盈余管理程度越高,公司财务危机发生的概率越高。（3）加入盈余管理程度变量能增加财务危机预警模型的解释能力。（4）加入盈余管理、公司治理结构特征的财务预警模型也能提高财务危机预警模型的解释能力,其预测的正确概率分别为前一年的98．9％,前二年为90．6％,前三年为75．3％。论文最后提出,如果上市公司能加强公司治理,进行盈余预测,可以很好的控制财务危机的发生概率。     

已婚女性规避家庭暴力风险的性别角色观念研究

依据全国妇联第三期中国妇女社会地位调查中个人调查表主卷湖南省部分数据,通过SPSS19.0软件进行已婚女性的性别角色观念状况及遭受配偶家庭暴力状况的单变量描述统计和双变量交叉分析。建立家庭暴力影响因素的多元线性回归模型,找出规避家庭暴力风险的变量因素,并通过二项Logistic回归分析已婚女性规避家庭暴力风险的性别角色观念因素。结果表明:性别角色观念对已婚女性规避家庭暴力有着积极的作用,较为现代化的性别角色观念能够有效降低其遭受配偶家庭暴力的风险。     

我国城市空气污染数据的准确性判别研究——基于我国东部三大城市群的比较分析

本文针对城市空气污染数据的真实性判别及分析问题,以多元线性回归、主成分分析、参数估计、相关分析为研究手段与方法,通过建立多元线性回归模型、各城市群气象数据真实性判定模型、双变量相关分析等模型,得出以下结论:(1)东部三大城市群中,京津冀城市群空气污染数据的准确度低于珠三角与长三角,这可能是由于京津冀极端天气情况较多造成的后果;(2)工业产值与空气质量指标AQI指数、PM2.5、PM10较强的相关性.     

基于Stacking集成学习的肺癌患者存活性预测模型研究

为提高肺癌患者存活性预测的准确率,提出一种基于Stacking集成学习的肺癌患者存活性预测模型.先对数据集进行预处理、特征选择、变量转换等,然后以XGBoost(eXtreme Gradient Boosting)、SVM(Support Vector Machine)和LR(Logistic Regression)3种算法为基学习器,以朴素贝叶斯为元学习器构造模型,再运用Grid Search网格搜索方法优化超参数,并利用交叉验证方法对SEER公开的肺癌数据集进行仿真实验.研究结果表明,该模型的预测准确率达85%,比单一模型高10%.该模型在肺癌患者存活性预测上有着更好的准确性和解释性,可以很好地为肺癌患者预后提供决策支持,以弥补经验的不足.     

浅谈Copula函数在金融风险分析中的应用

金融风险来自金融工具价格的变动,随着金融工具及其衍生产品的多元化发展,各种风险因素之间的关系日趋复杂和紧密,基于线性相关的金融分析模型已不能满足风险分析管理的需要。Copula函数具有许多优良的性质,可以将多元分布分解为单个变量的边缘分布和一个描述变量之间相关结构的Copula函数,从而简化模型的复杂性,提高模型的实用性和有效性。将Copula函数引入到金融风险分析中,可以更加准确的反映资产之间的相关结构,从而提高模型预测的准确性。     

基于光谱技术的牛肉嫩度等级无损判别

为实现生鲜牛肉嫩度等级的无损快速判别,用可见近红外光谱检测实验系统,获取不同嫩度的40块牛肉样品的400~1650 nm范围的光谱信息,以肉类嫩度测量仪测量牛肉剪切力值为分类依据.用多元散射校正(MSC)、SG平滑处理方法处理光谱数据,分别建立牛肉嫩度的线性判别分类模型和支持向量机分类模型.结果显示,用MSC+SG预处理后所建立的线性判别分类模型预测能力最好,训练集和测试的回判识别率和预测识别率分别为96.7%、100%.结果表明,用光谱技术结合线性判别分析,可以对牛肉嫩度进行无损快速判别.     

银黄颗粒剂的近红外光谱鉴别分析

利用近红外光谱与判别分析相结合的方法对不同生产厂家的银黄颗粒剂进行分类鉴别。采用多元散射校正(MSC)+一阶导数对银黄颗粒剂的近红外光谱进行预处理,在7400～4000cm-1光谱区间内建立近红外判别分析模型,并用三重交叉验证的方法对模型的稳定性进行验证。所建判别分析模型对校正集样品的分类鉴别准确率为100%,对验证集样品的预测鉴别准确率为100%。结果表明,该方法准确、快速、简便,可用于不同生产厂家银黄颗粒剂的分类鉴别。     

Comparing the Fit of Item Response Theory and Factor Analysis Models

Linear factor analysis (FA) models can be reliably tested using test statistics based on residual covariances. We show that the same statistics can be used to reliably test the fit of item response theory (IRT) models for ordinal data (under some conditions). Hence, the fit of an FA model and of an IRT model to the same data set can now be compared. When applied to a binary data set, our experience suggests that IRT and FA models yield similar fits. However, when the data are polytomous ordinal, IRT models yield a better fit because they involve a higher number of parameters. But when fit is assessed using the root mean square error of approximation (RMSEA), similar fits are obtained again. We explain why. These test statistics have little power to distinguish between FA and IRT models; they are unable to detect that linear FA is misspecified when applied to ordinal data generated under an IRT model.     

An empirical analysis of an alternative configuration of the Academic Motivation Scale

The purpose of this study was to examine an alternative configuration of the Academic Motivation Scale (AMS), a psychometric instrument used as a measure of academic motivation in various academic environments. The analyses utilised data from a convenience sample of 2354 business students, broken into two random subsamples of 1177 cases. Exploratory factor analysis of the 28 AMS scale items was conducted on the estimation sample. The results indicated good model fit for a four-factor configuration consisting of amotivation, external regulation, identified regulation, and intrinsic motivation as indicated by factor loadings, as well as internal consistency and reliability statistics.

Confirmatory factor analysis was conducted on the holdout sample to independently assess the construct validity and internal consistency of the item loadings on the reconfigured scale. Results indicated good model fit for the four-factor configuration and a significant loss of fit for competing three- and one-factor models. The uncovered factor structure advances our understanding of how the items on this scale cluster into theoretically meaningful constructs. This knowledge may be utilised in research designed to further assess the impact of motivational states on educational outcomes such as academic performance, absenteeism, dropout rates, etc.     

Scaling Performance Assessments: A Comparison of One-Parameter and Two-Parameter Partial Credit Models

In one study, parameters were estimated for constructed-response (CR) items in 8 tests from 4 operational testing programs using the l-parameter and 2- parameter partial credit (IPPC and 2PPC) models. Where multiple-choice (MC) items were present, these models were combined with the 1-parameter and 3-parameter logistic (IPL and 3PL) models, respectively. We found that item fit was better when the 2PPC model was used alone or with the 3PL model. Also, the slopes of the CR and MC items were found to differ substantially. In a second study, item parameter estimates produced using the IPL-IPPC and 3PL-2PPC model combinations were evaluated for fit to simulated data generated using true parameters known to fit one model combination or ttle other. The results suggested that the more flexible 3PL-2PPC model combination would produce better item fit than the IPL-1PPC combination.     

In Search of Optimal Cognitive Diagnostic Model(s) for ESL Grammar Test Data

This study compares five cognitive diagnostic models in search of optimal one(s) for English as a Second Language grammar test data. Using a unified modeling framework that can represent specific models with proper constraints, the article first fit the full model (the log-linear cognitive diagnostic model, LCDM) and investigated which model emerged as the dominant model. It then fit the dominant model and the other models to confirm that the model provides the best fit to the data. The model found to represent the most number of items in the test was the Compensatory Reparameterized Unified Model (C-RUM) and other models compared were the Deterministic-Input, Noisy-And (DINA), Deterministic Input, Noisy-Or-gate (DINO), and Noisy Input, Deterministic-Or-gate (NIDO). The absolute (item-association root mean square error values) and relative (information criteria) model fit indices also indicated that the LCDM and the C-RUM were the best fit to the data. More detailed analyses on the functioning of the C-RUM were conducted and the interpretation of the results was included in the discussion section. The article ends with some suggestions for future research based on the limitations of the study.     

Relative and Absolute Fit Evaluation in Cognitive Diagnosis Modeling

As with any psychometric models, the validity of inferences from cognitive diagnosis models (CDMs) determines the extent to which these models can be useful. For inferences from CDMs to be valid, it is crucial that the fit of the model to the data is ascertained. Based on a simulation study, this study investigated the sensitivity of various fit statistics for absolute or relative fit under different CDM settings. The investigation covered various types of model–data misfit that can occur with the misspecifications of the Q‐matrix, the CDM, or both. Six fit statistics were considered: –2 log likelihood (–2LL), Akaike's information criterion (AIC), Bayesian information criterion (BIC), and residuals based on the proportion correct of individual items (p), the correlations (r), and the log‐odds ratio of item pairs (l). An empirical example involving real data was used to illustrate how the different fit statistics can be employed in conjunction with each other to identify different types of misspecifications. With these statistics and the saturated model serving as the basis, relative and absolute fit evaluation can be integrated to detect misspecification efficiently.     

The Overlap of Child and Adult Writing Vocabularies

Many mechanistic rules of thumb for evaluating the goodness of fit of structural equation models (SEM) emphasize model parsimony; all other things being equal, a simpler, more parsimonious model with fewer estimated parameters is better than a more complex model Although this is usually good advice, in the present article a heuristic counterexample is demonstrated in which parsimony as typically operationalized in indices of fit may be undesirable. Specifically, in simplex models of longitudinal data, the failure to include correlated uniquenesses relating the same indicators administered on different occasions will typically lead to systematically inflated estimates of stability. Although simplex models with correlated uniquenesses are substantially less parsimonious and may be unacceptable according to mechanistic decision rules that penalize model complexity, it can be argued a priori that these additional parameter estimates should be included. Simulated data . are used to support this claim and to evaluate the behavior of a variety of fit indices and decision rules. The results demonstrate the validity of Bollen and Long’s (1993) conclusion that “test statistics and fit indices are very beneficial, but they are no replacement for sound judgment and substantive expertise” (p. 8).     

The equal correlation baseline model for comparative fit assessment in structural equation modeling

In this article I describe and evaluate an alternative baseline model for comparative fit assessment of structural equation models and compare it to the standard “null” baseline model. The new “equal correlation” baseline model constrains all variables to have equal, rather than zero, correlations, whereas all variances are free. The new baseline model reflects the reality of atheoretical background correlation in nonex‐perimental data sets, and it improves the ability of comparative fit indices to distinguish between better and worse target models. It also helps to preserve the statistical link between these indices and the noncentral χ² distribution. Also, computing the same comparative fit indices using different baseline models will provide more information about model fit than computing multiple comparative fit indices using the same baseline. I also point out some limitations of the proposed baseline model.     

How Often Is the Misfit of Item Response Theory Models Practically Significant?

Standard 3.9 of the Standards for Educational and Psychological Testing ( 1999 ) demands evidence of model fit when item response theory (IRT) models are employed to data from tests. Hambleton and Han ( 2005 ) and Sinharay ( 2005 ) recommended the assessment of practical significance of misfit of IRT models, but few examples of such assessment can be found in the literature concerning IRT model fit. In this article, practical significance of misfit of IRT models was assessed using data from several tests that employ IRT models to report scores. The IRT model did not fit any data set considered in this article. However, the extent of practical significance of misfit varied over the data sets.