Indonesian load prediction estimation using long short term memory

Erliza Yuniarti, Siti Nurmaini, Bhakti Yudho Suprapto


Prediction of electrical load is important because it relates to the source of power generation, cost-effective generation, system security, and policy on continuity of service to consumers. This paper uses Indonesian primary data compiled based on data log sheet per hour of transmission operators. In preprocessing data, detrending technique is used to eliminate outlier data in the time series dataset. The prediction used in this research is a long-short-term memory algorithm with stacking and time-step techniques. In order to get the optimal one-day forecasting results, the inputs are arranged in the previous three periods with 1, 2, 3 layers, 512 and 1024 nodes. Forecasting results obtained long short-term memory (LSTM) with three layers and 1024 nodes got mean average percentage error (MAPE) of 8.63 better than other models.


detendring; load forecasting; long short term memory; one-day forecasting;

Full Text:




  • There are currently no refbacks.

View IJAI Stats

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.