The concept of neural network applications to the analysis of weather parameters for risk prediction

The weather, especially its fluctuations, has a significant impact on the road driv- ing conditions - leads to a number of collisions and as a consequence to hundreds of deaths, thousands injured and millions of economic losses. The article briefly specifies the methods of particular weather elements description which can be found in literature and their impact on the road conditions. Additionally, weather parameters analysis was performed. On its ba- sis adaptation of a neural network to predict temperature changes and the possibility of icing was proposed. The tests and results obtained by using the designed neural network could be used in fuzzy expert systems.