A Comparative Study of Gaussian Process Machine Learning and Time Series Analysis Techniques for Predicting Unemployment Rate
This study explores and compares the capability of Gaussian process machine learning (GPML) with time series analysis techniques, which are autoregressive integrated moving average (ARIMA), seasonal ARIMA (SARIMA) and cubic spline interpolation, in modeling unemployment rate in Malaysia over the per...
| Published in: | 2024 16th International Conference on Computer and Automation Engineering, ICCAE 2024 |
|---|---|
| Main Author: | |
| Format: | Conference paper |
| Language: | English |
| Published: |
Institute of Electrical and Electronics Engineers Inc.
2024
|
| Online Access: | https://www.scopus.com/inward/record.uri?eid=2-s2.0-85198357359&doi=10.1109%2fICCAE59995.2024.10569432&partnerID=40&md5=e74843de516322667c0a3e37be73911f |
| Summary: | This study explores and compares the capability of Gaussian process machine learning (GPML) with time series analysis techniques, which are autoregressive integrated moving average (ARIMA), seasonal ARIMA (SARIMA) and cubic spline interpolation, in modeling unemployment rate in Malaysia over the period of 1991 to 2022. Six predictive models are developed based on the observations. The predictive performance of each model is quantified using mean absolute error (MAE) and mean squared error (MSE). GPML demonstrated the best predictive model, exhibiting minimal MAE and MSE compared to the time series analysis techniques. The statistical analysis also concluded no significant mean difference between the GPML model and the actual observations, implying a robust predictive model in predicting the unemployment rate in Malaysia. © 2024 IEEE. |
|---|---|
| ISSN: | |
| DOI: | 10.1109/ICCAE59995.2024.10569432 |