The article considers the problem of diagnostics of respiratory diseases. The authors have chosen as a possible direction of solution of the given problem neuronetwork convolution model, developed and showed its possible architecture. The article also shows what steps of data pre-processing must be produced for the best result of model training. The authors suggest that the diagnosis of respiratory diseases should be carried out by analyzing the characteristics of audio recordings of breathing. Such audio recordings are already being produced, making it possible for the network to use a data set that is freely available on the Internet. The paper contains a list of characteristics of audio recordings of breathing that can be used to analyze and make a preliminary diagnosis. An important part of the work is the analysis of existing modern scientific methods that allow somehow simplify the work of medical personnel and help to preserve the health of the patient. The given results of learning the neural network model show which diseases can be diagnosed with great confidence automatically, and which are more difficult to determine and require additional research.

No abstract

No abstract

No abstract

No abstract