Hledání shluků podobných dopravních nehod
Loading...
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Tato diplomova prace se zabyva hledanim nebezpecnych dopravnich useku v Ceske republice. K jejich nalezeni je vytvoren analyticky nastroj pro hledani shluku dopravnich nehod s moznosti zkoumat jakykoliv vstupni soubor s validnimi GPS pozicemi. V tomto nastroji si uzivatel muze zvolit z nekolika metod na hledani nebezpecnych shluku. Filtrovat data pro vstup do metody lze pomoci jakehokoliv atributu obsazeneho ve vstupnim souboru a podle vsech hodnot daneho atributu je automaticky vybran typ ltru, jako napriklad casovy nebo ciselny rozsah.Jelikoz se jedna o analyticky nastroj, je mozne kazdy nalezeny shluk analyzovat. Pri teto analyze jsou ve shluku hledany spolecne priciny nehody na vybranych atributech. Vystupem jsou pak asociacni pravidla, ktera slouzi k detekci predpokladu ke vzniku nehody v danem miste.
This diploma thesis work deals with the search for dangerous stretches of roads with high incidence of traffic accidents in the Czech Republic. An analytical tool has been created in order to search for areas of high traffic accident clusters. Any input file with valid GPS positions can be researched. Several methodS may be applied in order to search for these accident clusters. Data can be filtered for input to methods by any attribute in the file. A filter is selected automatically by values in the attribute. There are a few types of filters, for example, time range or numerical range.Any found cluster can be analyzed with this tool. This analysis is supposed to find common causes of traffic accidents in a cluster in selected attributes. The output of this analysis are association rules which serve to detect the assumption of the occurence of an accident in a selected cluster.
This diploma thesis work deals with the search for dangerous stretches of roads with high incidence of traffic accidents in the Czech Republic. An analytical tool has been created in order to search for areas of high traffic accident clusters. Any input file with valid GPS positions can be researched. Several methodS may be applied in order to search for these accident clusters. Data can be filtered for input to methods by any attribute in the file. A filter is selected automatically by values in the attribute. There are a few types of filters, for example, time range or numerical range.Any found cluster can be analyzed with this tool. This analysis is supposed to find common causes of traffic accidents in a cluster in selected attributes. The output of this analysis are association rules which serve to detect the assumption of the occurence of an accident in a selected cluster.
Description
Subject(s)
shlukova analyza, DBSCAN, OPTICS, DENCLUE, asociacni pravidla, Apriori, doprava, nehody, hustota dopravy, clustering analysis, DBSCAN, OPTICS, DENCLUE, association rules, Apriori, traffic, accidents, traffic density