Review of the scientific supervisor of baccalaureate thesis by Grigorieva Irina Vladimirovna.
"Canonical analysis of categorical data with application in marketing"

Despite the fact that the theme of bachelor work has the econometric character, from a mathematical point of view it covers several important issues of data analysis sections. Improvement of the developed methods relevant for other statistical applications.
On the one hand this is a problem of reduction of dimension for categorical data based on symptom-syndromic approach,  on the other hand to investigate the relationship between  two sets of categorical data, and the study of the structure multivariate linear combinations of symptoms over a finite field, which have approximately the same properties of low coefficient of uncertainty. To do this, it introduces the concept of nominative representative - a symptom that results in the greatest loss of information when removing it from the syndrome.  From the practical point of view, to investigate the relationship between the characteristics of innovation projects, which were observed at an interval of one year. Сlassical methods of statistical analysis were used to solve the problem (factor analysis and analysis of variance, independence criterion qualitative characteristics chi-square, statistic information: entropy and coefficient of uncertainty). The frequency method of search of the nominative representative is offered, it proved the assertion of the positivity of the difference between the uncertainty coefficients at the removal of one symptom from the syndrome and the equality of its zero in the case of independent features. Necessary programs have been developed. To accelerate the calculating has been studied and applied in practice the algorithm of fast enumeration Grassmannians points (Anan'evskaya P.V.) because the symptom-syndromic approach requires a lot of computer time. 
Thus, all the necessary skills have been demonstrated, previously unknown facts have been briefly told, evidence of new proposals have been presented, data and results have been structured and treatments have been detailed. Notes: the writing style is close to the report, typos and small study of some aspects, in particular, criterion information content of the component symptom. The rating "good".
Scientific supervisor, associate professor of the department of statistical modeling: Alekseeva N. P.