共查询到20条相似文献,搜索用时 406 毫秒
1.
Abimbola S. Young 《Higher Education》1989,18(3):321-339
Multi-group discriminant analysis is used to identify the pre-enrollment demographic and academic factors which best separate out the success classes in the first-year examinations of science students at a Nigerian university. Each science-based faculty is separately analysed to control for faculty differences in all the variables used. The pre-eminent predictor of success is performance in the specialist subjects of the universitys matriculation examination. Other variables selected for use in the discriminant functions depend on the faculty group. Misallocation rates are generally low and, exept for one faculty group, discrimination is uni-dimensional. 相似文献
2.
3.
田兵 《渭南师范学院学报》2014,(23):8-11
判别分析法是根据所研究个体的观测值来构建一个综合标准用来推断个体属于已知种类中哪一类的方法。 Fisher判别分析法是一种非常重要而且应用极为广泛的判别分析法。文章介绍了Fisher判别分析法的数学思想,详细阐述了在两个总体和多个总体情况下它的判别函数以及判别准则。之后通过举例说明了Fisher判别分析法在解决实际问题中的具体应用。 相似文献
4.
二维最大散度差线性鉴别分析方法不仅有效地避免了在人脸识别中传统的Fisher线性鉴别分析通常存在的“小样本问题”,而且使其特征抽取的速度有了大幅度的提高.本文通过引入著名的“核技巧”,将二维最大散度差线性鉴别分析扩展到非线性空间,提出了一种新的二维核最大散度差鉴别分析方法.该方法不仅抽取了图像中更加有效的非线性鉴别特征,使正确识别率显著提高,而且为二维非线性鉴别分析提供了一个统一的构架.最后在AR标准人脸库中的实验结果验证了本文算法的有效性. 相似文献
5.
胡煜 《广东技术师范学院学报》2010,(9)
所统计分析的数据集是前列腺癌基因数据集.采用分片逆回归方法和线性判别分析(LDA),二次判别分析(QDA).对基因芯片(微阵列)数据进行分析.用SIR降维,用LDA和QDA分类.讨论分片逆回归方法和二种方法对基因样本进行分类的效果. 相似文献
6.
Anthony JL Williams JM McDonald R Corbitt-Shindler D Carlson CD Francis DJ 《Annals of dyslexia》2006,56(2):239-270
Phonological awareness (PA), phonological memory (PM), and phonological access to lexical storage (also known as RAN), play
important roles in acquiring literacy. We examined the convergent, discriminant, and predictive validity of these phonological
processing abilities (PPAs) in 147 3-, 4-, and 5-year-old children whose native language was Spanish. Confirmatory factor
analysis (CFA) supported the validity of each PPA as separate from general cognitive ability and separate from each other.
Moreover, structural equation modeling found RAN uniquely associated with knowledge of Spanish letter names and Spanish letter
sounds. PA was found the best predictor of children’s ability to distinguish alphabetic text from nonalphabetic text. Finally,
general cognitive ability was only indirectly associated with emergent literacy skills via PPAs. These results highlight the
importance of PPAs in the early literacy development of native Spanish speaking preschool children. 相似文献
7.
Kernel Model Applied in Kernel Direct Discriminant Analysis for the Recognition of Face with Nonlinear Variations 总被引:1,自引:0,他引:1
A kernel-based discriminant analysis method called kernel direct discriminant analysis is employed, which combines the merit of direct linear discriminant analysis with that of kernel trick. In order to demonstrate its better robustness to the complex and nonlinear variations of real face images , such as illumination, facial expression, scale and pose variations, experiments are carried out on the Olivetti Research Laboratory, Yale and self-built face databases. The results indicate that in contrast to kernel principal component analysis and kernel linear discriminant analysis, the method can achieve lower (7%) error rate using only a very small set of features. Furthermore, a new corrected kernel model is proposed to improve the recognition performance. Experimental results confirm its superiority (1% in terms of recognition rate) to other polynomial kernel models. 相似文献
8.
《The Journal of educational research》2012,105(8):300-304
AbstractReports of all published factor analyses of student ratings of college faculty were reviewed to determine what common factors had emerged and to identify items likely to be useful in discrimination between teachers. A 39-item form was administered to students of 18 instructors both at the beginning and end of a semester course. Twenty-four items had been included in the form used earlier by Isaacson et al. (5). The current data and those collected by Isaacson et al. were analyzed by multiple discriminant analysis and the results were compared to those obtained by factor analysis. While results of the analyses were generally similar, the differences between the results of multiple discriminant analysis and factor analysis point to differences between student stereotypes of teacher behavior and differences in actual behavior between teachers. For uses in which the primary concern is to compare one teacher with another, the dimensions derived by multiple discriminant analyses seem likely to be more useful. 相似文献
9.
Marta Desimoni Teresa Gloria Scalisi Margherita Orsolini 《Learning and Instruction》2012,22(5):340-353
A sample of 170 Italian children was assessed for reading accuracy, reading speed, text comprehension and spelling in Grades 1 and 3 in order to investigate the concurrent and longitudinal relationships among literacy skills. Main results from multivariate analyses (regression, discriminant and path analyses) indicated that reading speed was the best predictor of later literacy and that spelling was the most stable measure and influenced text comprehension and reading speed. An asymmetry was also observed in the longitudinal relationship between reading and spelling errors, with reading errors predicting later spelling errors and a non-significant result in the opposite direction. 相似文献
10.
Self-Efficacy: An Essential Motive to Learn 总被引:9,自引:0,他引:9
Zimmerman BJ 《Contemporary educational psychology》2000,25(1):82-91
During the past two decades, self-efficacy has emerged as a highly effective predictor of students' motivation and learning. As a performance-based measure of perceived capability, self-efficacy differs conceptually and psychometrically from related motivational constructs, such as outcome expectations, self-concept, or locus of control. Researchers have succeeded in verifying its discriminant validity as well as convergent validity in predicting common motivational outcomes, such as students' activity choices, effort, persistence, and emotional reactions. Self-efficacy beliefs have been found to be sensitive to subtle changes in students' performance context, to interact with self-regulated learning processes, and to mediate students' academic achievement. Copyright 2000 Academic Press. 相似文献
11.
判别分析是多元统计分析的一种方法,已在物种分类、经济分析、地质勘探、天气预报等诸多领域得到广泛的应用。本文首先介绍了判别分析的数学模型和求解方法,然后结合第十次新增博士学位授予单位的实际评审材料,阐述了判别分析方法在此类评估中的应用,以期该定量分析方法能在教育评估领域得到有效的利用。 相似文献
12.
This study was undertaken to determine whether a biology preparatory course given at an urban community college was helping students to develop the proper skills and background necessary for them to successfully complete follow-up courses in biology. A group of students who enrolled in a biology preparatory course, and subsequently, a follow-up anatomy and physiology or general biology course (experimental group) was compared to a group of students who should have registered for the preparatory course, but who enrolled directly into the anatomy and physiology or general biology course (control group). It was shown that there was no significant difference in their anatomy and physiology or general biology grades. Furthermore, only 16% of the initial group of preparatory students enrolled in and passed a follow-up biology course. Examination of the preparatory group using discriminant analysis ascertained that mathematics score was the principle discriminator between pass/fail groups. A stepwise multiple regression analysis of the variables explaining the preparatory grade showed that mathematics score, reading score, and type of high school degree explained 33% of the variance. Of the students who did pass the preparatory course and enrolled in a follow-up biology class, their preparatory grade was a good predictor of their achievement (measured by follow-up course grade), as determined by multiple regression. 相似文献
13.
提出了一种新的基于核判别分析的手写汉字识别方法。核判别是对线性判别式分析的非线性判别分布的扩展。阐述了核判别分析法的基本原理,建立了核判别分析手写体识别模型,研究分析了核判别分析手写体识别模型的缺陷并提出了优化策略。在此基础上,采用C#与核判别分析相结合的算法,更好地展示了核判别算法的算法优势,采用高级语言提高了网络的学习训练速度和识别效果。 相似文献
14.
15.
Survey data from 458 former health science graduate students were examined for correlations between variables that might be used in selecting research training award recipients and successful training outcome defined as a highly productive research career. Scores on six research outcome variables were used to cluster respondents into two groups: active and less-active researchers. After exploring factors underlying thirteen predictor variables, discriminant analysis was employed to identify predictive strength. There were no significant direct or indirect correlations between posttraining research activity and either the overall undergraduate academic grade of respondents or the amount of encouragement they received from the sponsors of their graduate award applications. An additional seven variables also showed no significant correlation with research career outcome. They were science hobbies as a child; undergraduate prizes and awards; summers spent in research; undergraduate publications; prestige and research orientation of undergraduate environment; and prestige of graduate institution. Four variables that showed promise as useful indicators of posttraining research activity were investigative personality; research orientation of the graduate environment; focused energy personality; and, to a lesser extent, preuniversity interest in science. 相似文献
16.
The study examined the convergent and discriminant validity of three methods for assessing three subskills of reading: word analysis, vocabulary, and comprehension. These three subskills were measured by teachers' ratings, specialists' ratings, and standardized tests. Correlations of all three skills, each measured by the three different methods, were studied by the multi-trait-multimethod procedure. Although there was some support for convergent validity, the study revealed a total lack of discriminant validity for any of the three subskills of reading. 相似文献
17.
18.
商业银行客户信用评价判别分析模型及其应用 总被引:1,自引:0,他引:1
邓恩 《湖南广播电视大学学报》2005,(3):63-65
客户信用分析是商业银行风险管理的基础,而风险管理是商业银行取得经营成功的关键.本文用多元统计中的判别分析为方法,探讨了中国社会这个信用相对缺失的大环境下,如何实现对于商业银行客户贷款资信的综合评价. 相似文献
19.
胡煜 《鞍山师范学院学报》2007,9(4):20-24
主要用偏最小二乘(PLS)方法和二次判别分析(QDA)有监督分类的方法来对基因芯片(微阵列)数据进行分析.PLS是一种提取海量的数据有效特征的有效方法,可以获得与原来基因芯片数据更为接近的成分的提取特征的效果.结果表明用PLS方法事先对数据处理可以提高基因芯片数据分析的准确性. 相似文献
20.
Ravinder Koul Thanita Lerdpornkulrat Soontornpathai Chantara 《Journal of Science Education and Technology》2011,20(6):761-770
A student’s motivational orientation is considered to be a predictor of a range of related education decisions, from attending
classes to choosing a particular course or a profession. This survey study conducted with student volunteers (males = 519;
females = 904) enrolled in secondary school science-math academic stream in Thailand investigated the relationship between
measures of motivation (achievement goal orientation and physics and biology classroom anxiety) and aspirations for high earning
science and math related careers. Results of multiple discriminant analyses showed gender differences in the motivational
factors that influence career aspirations. Our interpretation of the findings highlights the significance of cultural beliefs
about gender in decision making for careers. 相似文献