Design and Development of a Discovery Service for Drivers Within the Connected Car Context Using Predictive Machine Learning Methods
-
Javier Goikoetxea [1] ; Alex Rayón [1]
-
[1] Universidad de Deusto
Universidad de Deusto
Bilbao, España
-
- Localización: Machine Learning and Artificial Intelligence with Industrial Applications: From Big Data to Small Data / coord. por Diego Carou Porto, Antonio Sartal Rodríguez, João Paulo Davim, 2022, ISBN 978-3-030-91005-1, págs. 193-208
- Idioma: inglés
- Texto completo no disponible (Saber más ...)
-
Resumen
Currently, cars are equipped with a large number of electronic sensors that fulfil several functions. These vary from receiving and issuing a signal, to allow the automation through the permanent exchange of data and information. However, in this last field the decentralization of suppliers and manufacturers, and the lack of fully connected car solutions, have limited the creation of new solutions for both the drivers at a microeconomic level and for the general safety at a macroeconomic level. The way in which companies have developed their services to tackle these challenges have been through business rules. Historically, companies mixed up the geolocation of the vehicle with the proposals of the businesses with their own interests. This was a product-oriented approach, rather than a driver-oriented approach we propose. Additionally, we propose the usage of machine learning techniques that could scientifically show which activations are better to improve the value proposals for the drivers. Considering this context, we present a discovery platform for the drivers that could permit the recommendation of a service or a product when needed with the final focus of saving money. We also identify which variables are the most important ones in the maintenance and usage of the car. Considering a wide variety of variables, we show which ones explain better the behaviour of the drivers and show them ways to save money accordingly.
