Methods
of identification of high-risk road sections used in Poland at present are
based on criterions that take
into account number of accidents, which not
always allows determining unsafe sections. What is more, the procedure of
seeking a road with a 100-meter step is often time consuming, especially when the
number of
collisions is also considered. An
algorithm of cluster analysis has
been elaborated, which makes it possible to adjust parameters of identification
process very easily. A researcher can input to the program the following
values: the length of analysis period, threshold length of identified road
sections, the number of accidents as well as the number of collisions. The
effectiveness of the algorithm has been verified in practice, giving
satisfactory results.