The identification of water that is fit for consumption and usage has proven to be a challenge. The parameters of water that are analyzed are typically Temperature, pH, turbidity, SAC etc.To develop an event detector to accurately predict changes in a time series of drinking water composition data, we selected the AE and RBM .