Intelligent Traffic Flow Prediction Using Optimized GRU Model

Facilitating citizens with accurate traffic flow prediction increases the quality of life. Roadside sensors and devices are used to capture live streams huge data Internet Things (IoT) is becoming popular for deployment effective Intelligent Transportation Systems (ITS). Traffic from datastreams require building a data-driven model. This challenging task has attracted researchers better interpretation characteristics. The core problem in modeling diversity trends unpredictable variations temporal dependencies. Initially, statistical shallow neural network models were applied some extent. Recently, deep learning come up proven promising outcomes. Gated Recurrent Unit (GRU) variation recurrent networks effectively prediction. Like other networks, GRU uses hyperparameters sliding window time-steps mechanism prepare tune Better tuning search optimal size tedious process. In this research work, we present an algorithm along steps optimization. Results obtained on real-time public dataset show higher capability proposed method reduce error average gain optimized model over untuned 4.5%. Furthermore, apply experiment that our approach improves accuracy stability.