The Prediction of Car Ownership with Machine Learning Approaches Nobel Bilimsel Eserler

The Prediction of Car Ownership with Machine Learning Approaches Nobel Bilimsel Eserler

Development in the range of vehicles has become a notable characteristic of modern towns, as the accumulation of individual wealth and the requirements for transport significantly motivate the ownership and utilization of vehicles. These developments have encouraged researchers to study in this field. Although several methods associated with car ownership forecasting have been described till now, there is a lack of hybrid methods for this topic. Therefore, the unique characteristic of this research lies in the integration of the Genetic Algorithm (GA) model with the Artificial Neural Network (ANN) as a hybrid ANN-GA. The Grey Model (GM) was also utilized to forecast independent variables from 2021 until 2040. Then, forecasting car ownership based on the best-developed network was conducted. In this regard, large-scale actual data from 1970–2020 was gathered, including population, GDP per capita, petrol price, and road length. A comparative analysis between ANN and hybrid models demonstrated the effectiveness of utilizing the hybrid ANN-GA approach concerning the best performance criteria. The results of this forecasting indicate that car ownership will be gradually increased by around 36% until 2040. These results can be utilized in specific, complex, or ambiguous environments because of the flexibility in several developed and developing countries.