Сравнительный анализ способов проведения факторного анализа на порядковых переменных
The paper considers different approaches to the factor analysis (FA) for ordinal data. In some studies it is necessary to find a latent variable behind the observed indicators measured on an ordinal scale. Classical factor analysis cannot be applied to those indicators as it is built on the Pearson correlation coefficient which is only applicable to interval variables. So the researcher faces a choice: to treat the ordinal variables as the interval ones, to dichotomize ordinal variables or to use special techniques for ordinal indicators such as replacing the correlation matrix or using Categorical principal components analysis (CatPCA). The study is based on a theoretical comparison of assumptions that underpin the algorithms of each applications and a statistical experiment and provides an answer to the question which of the above-mentioned factorization approaches is optimal for indentifying latent variables measured by ordinal indicators on a 3-point, 5-point or 10-point scale.
In this paper we improve the speed of the nearest neighbor classiﬁers of a set of points based on sequential analysis of high-dimensional feature vectors. Each input object is associated with a sequence of principal component scores of aggregated features extracted by deep neural network. The number of components in each element of this sequence is dynamically chosen based on explained proportion of total variance for the training set.We propose to process the next element with higher explained variance only if the decision for the current element is unreliable. This reliability is estimated by matching of the ratio of the minimum distance and all other distances with a certain threshold. Experimental study for face recognition with the Labeled Faces in the Wild and YouTube Faces datasets demonstrates the decrease of running time up to 10 times when compared to conventional instance-based learning.
The paper is focused on the problem of balanced information and communication technologies development in modern society and its measures. Existing indices allow estimating the progress of this development, but they are mostly constructed for between countries comparison. At the lower level, we face the problem of different data set structure that does not allow us to use same indices immediately. Another problem is a lack of measures that provide both the estimation of the overall progress in ICT and the progress in its particular aspects at the regional level. Additional attention should be paid to the possible increase of regional differentiation in order to prevent its transformation into extreme form of digital divide. We propose the model in the form of hierarchical system of indices that includes the advantages of information development measurement in particular aspects and as a whole. Sub-indices in the system are just orthogonal, and widely used principal component analysis is not functional for their aggregation. That is why a measure of closeness of the aggregate indicator to the best possible value is used to create the main index. The output oriented data envelopment analysis is used along with the aggregate indicator construction technique to compare the positions of the most advanced regions. We also introduce the method of measuring the differentiation in some aspects of information and communication development to assess the dynamics of this differentiation. The results of the study can be used to stimulate progress in information and communication development that prevents an excessive increase in regional differentiation.
The article gives an overview of the main approaches to verification of the internal rating methods of analysis of the creditworthiness. The article gives practical advice on the application of these techniques by an example of real internal rating technique.
This paper considers how to analyze the performance indicators of universities. The data of this study comes from Russian National Research Universities' statements that are available in open access on official web sites. The main purpose of this study is to define via factor analysis the most important indicators for ranking and performance auditing behind the constellations of performance measures. For this purposes Statistical Package for the Social Sciences (SPSS) was used. Then we conducted factor analysis by using the extraction method with principal component analysis to choose the main factors for rating and performance auditing. We suggest using the factor analysis result for evaluation of an auditee.
The article explores the procedural aspect of constructing structural and logical typologies with the aim of creating the innovation index - workers attitudes guiding innovation and innovation -related behavior at workplace.
Several approaches to the concept of fatherhood present in Western sociological tradition are analyzed and compared: biological determinism, social constructivism and biosocial theory. The problematics of fatherhood and men’s parental practices is marginalized in modern Russian social research devoted to family and this fact makes the traditional inequality in family relations, when the father’s role is considered secondary compared to that of mother, even stronger. However, in Western critical men’s studies several stages can be outlined: the development of “sex roles” paradigm (biological determinism), the emergence of the hegemonic masculinity concept, inter-disciplinary stage (biosocial theory). According to the approach of biological determinism, the role of a father is that of the patriarch, he continues the family line and serves as a model for his ascendants. Social constructivism looks into man’s functions in the family from the point of view of masculine pressure and establishing hegemony over a woman and children. Biosocial theory aims to unite the biological determinacy of fatherhood with social, cultural and personal context. It is shown that these approaches are directly connected with the level of the society development, marriage and family perceptions, the level of egality of gender order.