| Title |
Modeling of Artificial Neural Network Based on Big Data for the Prediction of Hourly Load in Summer Weekda |
| Authors |
Young-Jin Nam ; Ha-Hyun Jo |
| DOI |
http://dx.doi.org/10.5207/JIEIE.2019.33.12.049 |
| Keywords |
Artificial Neural Network; Big Data; Load Forecasting; Multi-Layer-Perceptron; Weather Information |
| Abstract |
In this paper, The MLP model was developed to predict the weekday of August during the 18 year summer season. The MLP model reflected a variety of factors affecting hourly loads: weather·Calendar information and big data. In addition, two MLP models were constructed, each using ELU and ReLU as an activation function. In order to evaluate the predictive performance, out of sample forecasting was performed. The results showed that the average MAPE for the ARIMA-X model as a benchmark was 3.88. However The MLP model 1 and The MLP model 2 was 2.69 and 2.49, respectively. so the artificial neural network model reflecting weather·calendar information and big data was more predictive performance than time series model |