Predicting parking space availability based on heterogeneous data using Machine Learning techniques
Mehmood, Hassan (2019-05-08)
Avaa tiedosto
Lataukset:
Mehmood, Hassan
H. Mehmood
08.05.2019
© 2019 Hassan Mehmood. Tämä Kohde on tekijänoikeuden ja/tai lähioikeuksien suojaama. Voit käyttää Kohdetta käyttöösi sovellettavan tekijänoikeutta ja lähioikeuksia koskevan lainsäädännön sallimilla tavoilla. Muunlaista käyttöä varten tarvitset oikeudenhaltijoiden luvan.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:oulu-201905101704
https://urn.fi/URN:NBN:fi:oulu-201905101704
Tiivistelmä
These days, smart cities are focused on improving their services and bringing quality to everyday life, leveraging modern ICT technologies. For this reason, the data from connected IoT devices, environmental sensors, economic platforms, social networking sites, governance systems, and others can be gathered for achieving such goals. The rapid increase in the number of vehicles in major cities of the world has made mobility in urban areas difficult, due to traffic congestion and parking availability issues. Finding a suitable parking space is often influenced by various factors such as weather conditions, traffic flows, and geographical information (markets, hospitals, parks, and others). In this study, a predictive analysis has been performed to estimate the availability of parking spaces using heterogeneous data from Cork County, Ireland. However, accumulating, processing, and analysing the produced data from heterogeneous sources is itself a challenge, due to their diverse nature and different acquisition frequencies. Therefore, a data lake has been proposed in this study to collect, process, analyse, and visualize data from disparate sources. In addition, the proposed platform is used for predicting the available parking spaces using the collected data from heterogeneous sources. The study includes proposed design and implementation details of data lake as well as the developed parking space availability prediction model using machine learning techniques.
Kokoelmat
- Avoin saatavuus [31941]