Modeling Power The GasDay model is the mathematical power behind the entire GasDay project. It combines artificial neural networks, multiple linear regression, and artificial intelligence to determine a forecast. Both the artificial neural network model and the linear regression model generate load estimates which are then displayed in the GasDay workbook. Finally, the dynamic post processor (the AI portion of the model) looks at both estimates as well as recent learned gas consumption trends, and generates the final load estimate.Over time, the GasDay model has been modified and improved. The model is trained using relevant historical data (GasDay draws upon more than 500,000 days of gas flow data), and it has a more accurate tracking of varying customer consumption patterns. The model also learns and adapts to new gas flow trends in real time, and it easily tracks growth in base load and heating load. Finally, the model has been programmed to make adjustments for the day of the week and special days, such as holidays. These improvements have helped the make the model even more accurate, and we continue to modify the GasDay model to better meet your needs. |
Mathematical Modeling
What Is GasDay? What Can GasDay Do For You? Our Commitment to Service Mathematical ModelingForecasting FAQ
