Forecasting energy consumption in Sweden
(2022) NEKH03 20221Department of Economics
- Abstract
- Machine learning has acquired a lot of attention in the economic forecasting literature in recent years. In this thesis we forecast Swedish energy consumption and compare the forecasting performance of a machine learning technique with that of more traditional time series models. In fact, the LSTM neural network is compared with ARIMA and VAR forecasts. We conclude that in our setting, while these newer techniques perform well under some conditions and are able to outperform the ARIMA forecast, they are not found to outperform the VAR model which remains the best modelling choice among those considered here.
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9085730
- author
- Ferdinand-Dreyfus, Jonathan LU
- supervisor
- organization
- course
- NEKH03 20221
- year
- 2022
- type
- M2 - Bachelor Degree
- subject
- keywords
- ARIMA, VAR, LSTM, Energy consumption, Machine Learning
- language
- English
- id
- 9085730
- date added to LUP
- 2022-10-10 09:12:03
- date last changed
- 2022-10-10 09:12:03
@misc{9085730, abstract = {{Machine learning has acquired a lot of attention in the economic forecasting literature in recent years. In this thesis we forecast Swedish energy consumption and compare the forecasting performance of a machine learning technique with that of more traditional time series models. In fact, the LSTM neural network is compared with ARIMA and VAR forecasts. We conclude that in our setting, while these newer techniques perform well under some conditions and are able to outperform the ARIMA forecast, they are not found to outperform the VAR model which remains the best modelling choice among those considered here.}}, author = {{Ferdinand-Dreyfus, Jonathan}}, language = {{eng}}, note = {{Student Paper}}, title = {{Forecasting energy consumption in Sweden}}, year = {{2022}}, }