(61r) Clustering-Based Forecasting Framework for the Energy Sector

Energy sector is increasingly becoming more decentralized, de-fossilized, and digitalized. This also means that integrating green energy sources will bring some challenges that need to be handled with smart applications. In this context, load forecasting has gradually become the central and integral process in the planning and operation of electric utilities, energy suppliers, system operators, and other market participants (Weron et al., 2006). That's why statistical analysis and educated guesswork like machine learning (ML) and deep learning (DL) are widely applied and tried to improve by academia and companies in today's deregulated energy market.

Energy demand/consumption data represents a challenge due to having non-linearity and volatility, which make it highly vulnerable to changes in exogeneous factors. In addition, a forecasting model may not completely capture and simulate the unique characteristics, leading to poor forecasting performance (Pradeepkumar et al., 2017). Indeed, each forecasting technique has benefits and drawbacks, and no single approach consistently outperforms the others for any given data set (Makridakis et al., 2018, Petropoulos et al., 2022). This makes exploring and comparing different forecasting methods essential to identify the most effective approach. In this respect, a clustering-based forecasting framework is proposed as a forecasting framework to predict the demand for energy products. With the proposed approach, different clustering algorithms will be applied to group similar data together, and then based on each cluster characteristic, different forecasting methods, including statistical, ML, and DL will be used to have more accurate predictions. The performance of the proposed clustering-based forecasting framework will be compared to traditional forecasting techniques to analyse its effectiveness.

References

1. Weron, 2006, Modeling and Forecasting Electrcity Loads and Prices: A Statistical Approach.
2. Makridakis, S. Spiliotis, V. Assimakopoulos, 2018, Statistical and Machine Learning forecasting methods: Concerns and ways forward, PloS One, 13, e0194889.
3. Petropoulos, D. Apiletti, V. Assimakopoulos, M. Z. Babai, D. K. Barrow, S. B. Taieb, C. Bergmeir, R. J. Bessa, J. Bijak, J. E. Boylan, and others, 2022, Forecasting: theory and practice, International Journal of Forecasting.
4. Predeepkumar, V. Ravi, 2017, Forecasting financial time series volatiliy using swarm optimization trained quantile regression neural network. Applied Soft Computing, 58, 35–52.