A Stacked Long Short-Term Memory Neural Networks for Parking Occupancy Rate Prediction

Abstract

In recent times, it is noticed that the number of vehicles is considerably increasing on the road. To locate the parking lot in cities is becoming a tedious job for drivers. To accurately predict the parking occupancy, we constructed a Stacked LSTM (Long short-term memory) model. This model can analyze sequential data precisely as it has a memory capacity. Stacking the LSTM hidden layers intensifies the model and increases the prediction accuracy. We validated the proposed predictive model in terms of key error indicators (RMSE and MAE). We also evaluated the proposed model with baseline time series methods like ARIMA and SARIMA. The Stacked LSTM model outperforms the traditional models and enhanced prediction accuracy.