Taxi passengers demand prediction using deep learning

Recently, Deep Learning using Tensor Flow has been activated, and various studies have been carried out to combine big data with Deep Learning. In the case of taxi-related research, research has been conducted mainly on generating profits by providing demand-responsive services starting with users' calls. The purpose of this study is to implement demand forecasting service by predicting passenger demand of Taxi by conducting Supervised Learning based on Taxi tachometer data, building data, and smart card data. Supervised Learning is implemented based on the getting on point of Taxi in order to implement the taxi demand predictive service. Accuracy is secured by learning building data and public transportation getting on / off data in Taxi data implemented through Supervised Learning. We could confirm the utilization of taxi demand predictive service through taxi passenger demand forecasting using the supervised learning of Deep Learning.