50+ days of Machine Learning

 Hello :) Today is Day 155! A quick summary of today: Real-time Prediction of Taxi Demand Using Recurrent Neural Networks Graph WaveNet for Deep Spatial-Temporal Graph Modeling GMAN: A Graph Multi-Attention Network for Traffic Prediction DNN-Based Prediction Model for Spatio-Temporal Data The updated graph with reference connections from Obisdian is: Real-time Prediction of Taxi Demand Using Recurrent Neural Networks Introduction This paper proposes a real-time method for predicting taxi demands in different areas of a city. A big city is divided into smaller areas and during a pre-set period of time, the number of taxi requests in each area is aggregated. This way taxi data becomes a sequence and a LSTM is applied. LSTM is capable of learning long-term dependencies by utilising some gating mechanisms to store information. Therefore, it can for instance remember how many people have requested taxis to attend a concert and after a couple of hours use this information to predict that the