KULESHOVA, A. Statistické zpracování atributivních dat o zmetkovitosti procesu [online]. Brno: Vysoké učení technické v Brně. Fakulta strojního inženýrství. 2023.

Posudek vedoucího

Hrabec, Pavel

This thesis deals with statistical processing of attributive data from a manufacturing process. The thesis consists of four main chapters, where the first two describe the statistical apparatus used, the third chapter describes studied process, and the fourth chapter applies described statistical apparatus to the collected data. The data used are thoroughly described, including the necessary preprocessing. The hypothesis tests described always stratify the data according to one of the recorded variables (product type, product subtype and responsible foreman) and then compare the data according to the observed average mistakes rate per “standardized” unit of said product. U-charts are then constructed for categories of subtypes with the same average mistakes rate to check the stability of the process and identify outlying observations. At the end of the chapter, Pareto analysis was performed with a deeper division of the described errors to identify the most common errors in the manufacturing process. Although the results presented are valid, relevant and suitably interpreted, the thesis is to be faulted in particular for its graphical layout (uneven size and font of figures and tables, occasionally overflowing outputs, etc.). Even though the author cannot be denied great diligence and commitment during our collaboration for this diploma thesis, the time available (since January this year), combined with the late request for anonymization of the data used, was too restrictive to submit a truly excellent work, which in my opinion the author is certainly capable of. I recommend the resulting thesis for defense and give it an overall grade of C.

Posudek oponenta

Popela, Pavel

The submitted thesis has successfully fulfilled the relatively broadly formulated requirements and objectives of the assignment at a very good level. The thesis has 52 pages including appendix. The text is divided into six main chapters. The procedure and scope of the solution can be considered adequate. The choice of the used methods was certainly positively influenced by the professional erudition of the supervisor, who recommended an original strategy for processing the industrial data. The author then proceeded responsibly to its implementation and achieved the presented results. The author has obviously used her practical experience in an industrial enterprise and brought an interesting problem from the real world. She got the support of her industrial chairmen for the use of statistical data from practice. In the data processing, she followed  the policy suggested by her supervisor and she developed and applied it in details. The author interprets the obtained results in an original way, and so, she proves a very good orientation in the used statistical procedures. The diploma thesis is oriented towards the use of the results in  practice. From the theoretical point of view it is an application of advanced statistical methods in the field of quality management. The author had to get acquainted with them beyond the scope of courses of her studies, and this deserves to be appreciated. The author has had to deal with some obstacles in the process of thesis writing, especially, with the quality of the available input industrial data, which, as it happens, was beyond her control. This seems to somehow limit the positive impact of the obtained results in this practical application. The arrangement of the thesis is logical and follows the above mentioned data processing policy. Occasional short one-sentence paragraphs are distracting. Some of the statements seem too bold (see, e.g. 2.6). The formulas should have been edited rather consistently using word processing tools. The graphic level is very good. Some of the formulas and images are obviously taken using printscreen, but when enlarged they become out of focus. The choice of capital letters in the chapter headings of the table of contents, together with the choice of fonts and spacing is quite striking. The thesis contains too many unnecessary and unnoticeable typos, which should have been removed at the latest during the final check. For example: "i" instead of "I" in the acknowledgements, Parreto in the conclusion etc. Even without giving specific examples, I would also like to express my doubts to the questionable word order in a number of sentences (see, for example, the introduction). The work with literature is at a very good level. In some parts I would have welcomed even more citations (see Introduction, 2.2, etc.). The list of references could have been handled in a uniform way.  In conclusion, regarding the final content and form of the thesis, I would like to appreciate how the student has persistently mastered her statistical knowledge. I would like to add with a smile that in my opinion she also permanently applied the advice of her supervisor and his statistical Monday group in a "hawkishly" (sokolsky) organized manner. In  contrast, we can also guess the influence of insights provided to her by a fellow statistician in a whirlwind of flamenco. I recommend the submitted thesis for defence.

eVSKP id 151288