Proposal of a Machine Learning Approach for Traffic Flow Prediction

Area of Science:

• Urban planning and intelligent transportation systems.
• Data science and machine learning applications.
• Environmental science and sustainable development.

Background:

• Global urbanization intensifies urban population growth, exacerbating transportation management challenges.
• Persistent traffic congestion, pollution, and safety risks impede urban progress despite mitigation efforts.
• Accurate traffic flow forecasting is identified as a key solution for urban traffic congestion.

Purpose of the Study:

• To address the challenge of predicting urban traffic flow.
• To propose a novel two-level machine learning approach for traffic flow forecasting.
• To enable traffic flow prediction in urban areas, including those without existing sensor infrastructure.

Main Methods:

• A two-level machine learning methodology is proposed.
• The first level utilizes unsupervised clustering to identify patterns in sensor data.
• The second level employs supervised machine learning models for prediction.

Main Results:

• The approach successfully extracts patterns from sensor data.
• It enables traffic flow prediction in urban environments.
• The model's predictive capability was validated in a real urban scenario.

Conclusions:

• The proposed two-level machine learning model offers an effective solution for traffic flow forecasting.
• This method can provide valuable insights for urban traffic management, even in sensor-scarce areas.
• Accurate traffic prediction is vital for mitigating urban congestion and improving city mobility.