ML models show difficulty forecasting volatile Australian electricity prices

A new study benchmarks six machine learning models for short-term electricity price forecasting in Australia's National Electricity Market. The research highlights significant challenges due to high price volatility, irregular patterns, and structural changes in the market. Tree-based models like GBRT demonstrated superior performance in price prediction compared to LSTMs and SVR, achieving an R-squared value of 0.88, though overall prediction accuracy remains low with high error rates. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

IMPACT Highlights the difficulty of applying current ML models to volatile energy markets, suggesting hybrid models and data augmentation for future improvements.

RANK_REASON Academic paper published on arXiv detailing machine learning model performance for a specific forecasting task.

Read on arXiv cs.LG →

• Australia
• National Electricity Market
• AWMLSTM
• CatBoost
• GBRT
• LSTM
• LightGBM
• SVR

• paper
• other