Stress urinary incontinence (SUI) is is not life-threatening disease, but affects the quality of life, significantly reducing it [1]. According to the world literature, the prevalence of SUI among the female population of Europe and America ranges from 34% to 38%. The disease 
In this regard, more and more attention is paid to the prevention of the development of SUI. Currently, there are no valid strategies for the prevention of stress incontinence, which makes it very urgent to develop measures aimed at this.
Recently, one of the trends and a key aspect of the development of medicine has become the introduction of artificial intelligence into the diagnostic process to help specialists from various fields [2,3].
The aim of the study was to assess the possibility of using artificial neural network (ANN) modeling in predicting the development of SUI in women.

Based on literature data, we have identified nineteen risk factors were that are important in the diagnosis of SUI in women.
The list of factors contains: heredity, body mass index, number of pregnancies, number of births, fetal macrosomia, operations on the pelvic organs, chronic constipation, chronic obstructive pulmonary disease, urinary tract infections, regular weight lifting, dyshormonal disorders, pelvic organ prolapse, presence complications during pregnancy, heart valves prolapse, varicose veins and aneurysms, myopia, lens  subluxation or flattening, duodenal dyskinesia, nephroptosis, hernia.
Thus, the final neural network contains, excluding the hidden layers, 20 neurons: 19 input and 1 output. A medical specialist needs to fill in all the above factors, and then obtain a prediction result: the presence or absence of a diagnosed disease.
The training process of a neural network is precedent-based.
To organize such a method of training, a medical specialist interviewed 58 patients: 37 patients with a diagnosed disease, and 21 patients disease-free.
The survey data was collected in the form of  .csv table, which was then loaded into the program and converted to the DataFrame format using the “pandas” library, after which the data was normalized to a single form from 0 to 1.
After preparing the data, they were divided into two parts: the training sample, which included 80%, and the testing sample, which included the remaining 20% of the data.
During the training, several methods were tested, in particular: the logistic regression method, the SVC method, the random forest method, the Gauss method, and the gradient descent method

According to the data obtained, the most effective training method was the gradient descent method, with the help of which it was possible to achieve a prediction accuracy of more than 80%.
In the final analysis of the data, the most important factors are chronic constipation, nephroptosis, hernia, heredity, chronic obstructive pulmonary disease.

As a result of our study, among all the above risk factors, the following were the most significant: chronic constipation, nephroptosis, hernia, heredity, chronic obstructive pulmonary disease.
Considering the lack of an adequate understanding of the independent risk factors in each case, associated with SUI and the development of this disease, ANN is an excellent method for further study of potential risk factors and the development of SUI.

Considering this potential advantage, the use of ANNs to understand the complex relationships between explanatory variables and outcome is warranted. Continuation of the study is required with an increase in the number of participating patients.
Our study demonstrates the ability to train and use ANNs to predict SUI using independent variables. ANNs can potentially play a role in making decisions on the development of preventive measures aimed at preventing the development of SUI.

Kastelein, A.W., Dicker, M.F.A., Opmeer, B.C. et al. Innovative treatment modalities for urinary incontinence: a European survey identifying experience and attitude of healthcare providers. Int Urogynecol J 28, 1725–1731 (2017). https://doi.org/10.1007/s00192-017-3339-y
Checcucci E., Autorino R., Cacciamani G. Artificial intelligence and neural networks in urology: current clinical applications // Minerva Urol Nefrol. 2020;72(1): 49-57.  doi: 10.23736/S0393-2249.19.03613-0. Epub 2019 Dec 12
Dreiseitl S., Ohno-Machado L. Logistic regression and artificial neural network classification models: a methodology review // J Biomed Inform. 2002; 35(5-6):352-9.  doi: 10.1016/s1532-0464(03)00034-0.