Fig. 8From: FML-kNN: scalable machine learning on Big Data using k-nearest neighbor joinsPersonalised hourly water consumption prediction. The predicted values follow the trend of the real ones, but miss sudden events. This is apparent during 16 and 20 h for #user1 and during 10 and 15 h for #user3. #user2 had very little consumption during that day and our framework predicted that no consumption will occurBack to article page