The aim of this work was the design and development of a prediction framework to forecast the occupancy levels of bike stations for different prediction horizons. The study objectives are threefold:
- To build a prediction model learned from a training dataset for automatically forecasting the occupancy levels of bike sharing stations;
- To improve the accuracy of such model by properly tuning the configuration parameters;
- To develop a client-server system and Android application to visualize the results of prediction;