Review report on Zhang Peichi Master Thesis titled
“Research on Diabetes Prediction Model Based on Machine Learning from   the Perspective of Cooperative Game Theory”.

The Master thesis of Zhang Peichi is a pioneering work devoted to the problems connected with mathematical diabetes prediction models. These models are based on machine learning and use the basic ideas of cooperative game theory. The proposed diabetes prediction model is based on one-class SVM. The Shapley Values and Tau-value are computed using the ideas of cooperative game theory. It is important to mention that Tau-value is the first time used in this type of models. This is a nontrivial achievement of the author. The computational experiment was done based on real data got from China medical institutions. The thesis is of great practical importance and the proposed approach can be used for many other illnesses not only in China, but also in our country.

But there are some important comments which I have to mention and which did not allow me to give the maximal mark. The important problem is that it seems that the author used the mathematical tools taken from the cooperative game theory without deep understanding and in more technical way. For instance when I discussed with him the use of the Shapley Value and the Tau Value he was unable to explain the  form of the characteristic function which is the most important tool in the  corresponding mathematical models. Also he did not care very much to understand deeply the used mathematics and the time he spent during the preparation of the thesis was not used in the optimal way.  
I evaluate the thesis as “good” – “C”, but I will be not against if the commission will decide to give the higher mark.

Supervisor,
Prof. Dr.Sci L.A.Petrosyan
27.05.2024.