Real-Time Validation of Retail Gasoline Prices

We provide a method of validating current gasoline (petrol) prices for a crowd-sourced app that primarily provides current gasoline prices. To validate prices reported by users of the app, we propose an approach that validates each price report in real time as it is entered by a consumer by comparing it to the current prediction of what the price is expected to be at the specified store at the present time. To do so, a forecast model is used to predict, with high accuracy, a price range for each store in real-time so that when a price is entered by a consumer it can be flagged if it falls outside the predicted range. We present the first experimental results concerning predicting the current price in real time at all stores in a city. The results indicate that there is a significant improvement in reducing the forecast error when using our Price Change Rules model over the modified Most Common Action model.