Supervised automatic evaluation for summarization with voted regression model

The high quality evaluation of generated summaries is needed if we are to improve automatic summarization systems. Although human evaluation provides better results than automatic evaluation methods, its cost is huge and it is difficult to reproduce the results. Therefore, we need an automatic method that simulates human evaluation if we are to improve our summarization system efficiently. Although automatic evaluation methods have been proposed, they are unreliable when used for individual summaries. To solve this problem, we propose a supervised automatic evaluation method based on a new regression model called the voted regression model (VRM). VRM has two characteristics: (1) model selection based on ‘corrected AIC’ to avoid multicollinearity, (2) voting by the selected models to alleviate the problem of overfitting. Evaluation results obtained for TSC3 and DUC2004 show that our method achieved error reductions of about 17–51% compared with conventional automatic evaluation methods. Moreover, our method obtained the highest correlation coefficients in several different experiments.     

线性模型中相关附加信息全部损失的条件(英文)

通过讨论回归系数的混合估计和Bayes估计的相关附加信息全部损失的条件,本文给出了线性模型中选择相关附加样本应遵守的一个原则,并研究了两个应用问题。     

从女子标枪成绩的多元回归分析中看样本含量与F 的取值对结论的影响

以女子标枪6 个变量因素为例, 对三组不同样本含量和F 的取值进行了多元回归分析, 结果表明: 在样本的成绩范围内, 样本n = 10 建立的“最优”回归方程与样本n = 60 , n = 30 的结论基本相同. 认为用样本n = 10建立“最优”回归方程既方便又省力. 强调在作逐步回归分析时, F 的检验临界值要取低些, 并指出: 在指标因素过多时, 应先聚类, 然后分别选出影响最大的指标建立“最优”方程.     

Regression through the Origin

This article describes situations in which regression through the origin is appropriate, derives the normal equation for such a regression and explains the controversy regarding its evaluative statistics. Differences between three popular software packages that allow regression through the origin are illustrated using examples from previous issues of Teaching Statistics.     

千山决明子总黄酮含量的测定

利用分光光度法，测定了千山决明子总黄酮的含量，建立了芦丁浓度—吸光度关系的回归方程。     

An Excel Solver Exercise to Introduce Nonlinear Regression

Business students taking business analytics courses that have significant predictive modeling components, such as marketing research, data mining, forecasting, and advanced financial modeling, are introduced to nonlinear regression using application software that is a “black box” to the students. Thus, although correct models are estimated, students often do not obtain a thorough understanding of the nonlinear estimation process. The exercise presented in this article was created to demonstrate to students the need for nonlinear regression estimation—rather than using linear transformations and Ordinary Least Squares (OLS) and subsequently demonstrate the nonlinear optimization process to estimate nonlinear regression models. Using the spreadsheet exercise, students can see effects on the fit of the model by changing the model parameters as they change the values of the decision variables. After applying the spreadsheet to further exercises, students have expressed a deep understanding of the linear regression software. This exercise is innovative because the active learning exercise requires the students to make the logical connections between the structure of the model, the model's parameters, and the objective function.     

基于最小二乘支持向量机的油罐故障诊断方法

根据油罐故障分析,建立了油罐故障诊断模型。采用新型的支持向量机-最小二乘向量机(LS-SVM)的算法对此诊断模型进行预测,获得了满意的效果。该方法易于使用,很少受不确定性因素的影响,有很高的预测准确性。     

Assessing Longitudinal Change: Adjustment for Regression to the Mean Effects

Pascarella (J Coll Stud Dev 47:508–520, 2006) has called for an increase in use of longitudinal data with pretest-posttest design when studying effects on college students. However, such designs that use multiple measures to document change are vulnerable to an important threat to internal validity, regression to the mean. Herein, we discuss a brief history of regression to the mean and illustrate a straightforward procedure to make adjustments to initial pretest scores for regression to the mean effects utilizing a method developed by Roberts (in: G. Echternacht (Guest ed.) New directions for testing and measurement, 1980). Analyses are shown with both unadjusted and adjusted pretest scores, illustrating dramatic differences in conclusions about whether students change across time.

我国高技术产业竞争力与技术创新的关系研究

伴随科学技术的迅猛发展以及知识经济的兴起,技术创新已成为推动产业发展和经济增长的不竭动力。在设计了技术创新能力和产业竞争力指标体系的基础上,以我国高技术产业2003—2008年28个省市(西藏、青海、新疆除外)的面板数据为样本,运用回归分析的方法对技术创新与产业竞争力之间的关系进行了实证研究,深入分析了影响产业竞争力的技术创新关键要素。实证结果表明:技术创新能力与产业竞争力有显著的相关关系,其中创新研发能力和创新间接产出显著影响产业的效益竞争力,创新研发能力还显著影响产业的成长竞争力。     

基于水环境质量控制的高集约化农业景观格局优化研究

农业非点源污染治理已经成为制约高集约化农业区可持续发展的瓶颈之一。本文以洞庭湖高集约化农业区金井镇为案例,在分析景观格局和水污染特点的基础上,采用多元回归法建立景观格局指数与水环境质量季节关系模型,基于模拟结果分别从景观格局优化、生活废水和固体废弃物处理、养殖业调控出发,提出农业景观格局优化模式。评价结果表明：优化后水环境质量提升,总氮浓度控制在地表水环境质量最低标准内,农业直接经济效益和生态系统服务价值均高于基准年。本文提出的模型构建方法不仅验证了景观格局与水环境质量关系,也为农业景观格局优化提供依据,其结果可为政府制定农业可持续发展、生态环境保护等提供科学依据。