To understand the impact of new pricing structure on residential electricity demands, we need a baseline model that captures every factor other than the new price. The standard baseline is a randomized control group, however, a good control group is hard to design. This motivates us to develop data-driven approaches. We explored many techniques and designed a strategy, named LTAP, that could predict the hourly usage years ahead. The key challenge in this process is that the daily cycle of electricity demand peaks a few hours after the temperature reaching its peak. Existing methods rely on the lagged variables of recent past usages to enforce this daily cycle. These methods have trouble making predictions years ahead. LTAP avoids this trouble by assuming the daily usage profile is determined by temperature and other factors. In a comparison against a well-designed control group, LTAP is found to produce accurate predictions.
}, year = {2017}, journal = {International Journal of Big Data Intelligence (IJBDI)}, volume = {Vol. 5}, pages = {3-20}, month = {09/2017}, url = {http://www.inderscience.com/info/inarticle.php?artid=88269}, doi = {10.1504/IJBDI.2018.10008133}, note = {This is the preprint version of a paper published in International Journal of Big Data Intelligence. The published version of the article can be found here:
http://www.inderscience.com/info/inarticle.php?artid=88269
doi:10.1504/IJBDI.2018.10008133
}, language = {eng}, }