Artificial intelligence predicts high electricity bills before they arrive
By analyzing two years of historical data, a new system identifies households likely to see a sudden 50 percent spike in power costs.
Predicting a household's monthly energy costs usually requires waiting for the meter reading, but a new system uses historical patterns to forecast expenses mid-cycle. By training on two years of a customer's power usage data, the technology can estimate the final monthly bill just ten days after a billing cycle begins.
Continue Reading in App
2 more paragraphs · plus a 3-question quiz