Analysis and Prediction of Security Factors in Public Passenger Transport Using Artificial Neural Networks

Abstract :

This paper sets out to provide for an analysis and prediction of selected security factors in public passenger transport based on the example of the capital city of Warsaw.? An analysis of the literature is made, to point to selected studies on the subject. Then, the most common criteria for assessing public passenger transport found in the literature are presented. Taking into account the standard quality guidelines, [12] including for public passenger transport safety, as well as the criteria suggested by other authors listed in the first part of the paper, the factors were selected for analysis. Data used for the purposes of this paper were collected on the basis of a statistical guide kept by the Public Transport Authority in Warsaw. The time range of the study covers the period from January 2017 to December 2019.??Artificial neural networks (time series) were used for predictions. The purpose of time series analysis using ANN is to predict future values of a variable or variables, based on previous values of the same or another variable or variables (dependent variable values may go beyond the data range). The range was set for 12 months. A total of 320 network designs were trained for 8 selected factors, of which 5 best were selected for each factor and a prediction was made for a period of 12 months.? The quality of the neural network learning process was obtained at >80%. On this basis, the development trends for selected factors in the near future were demonstrated.