@article{blcian2023economic,
  title = {An Economic Analysis of Neural - Based Techniques Applied for Energy Consumption Prediction},
  author = {Delia BĂLĂCIAN and Stelian STANCU},
  year = 2023,
  url = {https://ibimapublishing.com/articles/JEERBE/2023/557970/},
  journal = {Journal of Eastern Europe Research in Business and Economics},
  volume = 2023,
  pages = 12,
  doi = 10.5171/2023.557970,
  abstract = {Energy consumption is of tremendous importance in modern society because of its impact on both climate change mitigation and the economy. In order to develop methods for reducing energy loss and carbon emissions, research has focused on improving building energy monitoring and planning. To achieve climate neutrality by 2050, the European Commission has set ambitious targets for increased energy efficiency in buildings and the increased use of renewable energy sources. The stakes of having robust prescriptive mechanisms to support the usage of renewable energy have grown once again in the context of the recently announced action plan REPowerEU. The research analyzes a dataset with over 8,000 entries and 123 variables, 14 of which were collected from an individual building and the rest were derived from the dataset by calculating the minimum, maximum, mean, standard deviation, hour, day, month, and wind direction. The case study compares the best models from two architectures of Artificial Neural Networks, Feed Forward and Long Short-Term Memory, in terms of performance by calculating the values for R-Squared, Mean Absolute Error and training time. In addition, comparisons between Adam and Adagrad as weight correction optimization functions are made. The key findings of the study indicate that the two types of architectures have similar performance metrics but Feed Forward Neural Network models are preferred to the Long Short – Term Memory ones given that the training time is significantly shorter while the R-Squared value is similar. The analysis continues with looking at the learning curves for each optimization function, unraveling that Adagrad is a better option for both types of neural networks than Adam.},
  keywords = {Energy Consumption, Artificial Neural Networks},
  note = Article ID: 557970
}