Computer Remote Business Management System

This thesis qualification work relates to the actual problem of the introduction of electronic document management using the functions of artificial intelligence and machine learning in order to predict sales volumes. The scientific novelty of the study consists in the fact that for the first time the dependence of the accuracy of time series prediction on the number of eras of machine learning of a neural network based on historical data was studied. The practical significance of the work lies in the fact that the model that is optimal in terms of time and quality of forecasting will make it possible to improve the quality and speed of making time series forecasts. The purpose of the work is to develop a remote business management system with a workflow automation system and the ability to forecast sales volumes for certain business goods. The study analyzed existing methods for designing web applications, developing electronic document management systems, processing and storing information, methods for predicting time series and assessing the accuracy of forecast data. During the study, algorithms in the Python programming language were developed and used to process historical data, build a model of a neural network, calculate its optimal parameters by calculating the coefficient of determination and the disparity coefficient of Teyl, and constructing forecast graphs. The computer system was tested on several practical data of different dimensions for all scenarios of the web application. The work has a practical implementation, published on the Internet with the url: http://151.248.125.119/index/. The designed system can be used by large stores and enterprises that have a need to translate documents into a virtual environment, automate the processes of employees of all positions, there are historical data on sales of their products for 2-3 years.