Entries for Net Load Forecasting Prize

We will be using probabilistic models for energy forecasting using diffusion based Variational Autoencoder (DVAE)

Developing solar-powered communities for aviation fixed based operations and urban areas. Harnessing sustainable energy for a greener future

We fit a semiparametric latent factor model with variational inference to provide efficient, well-calibrated predictions for future loads.

Integrate observation and numerical weather prediction with machine learning models for physics-informed probabilistic net load forecast

We are looking at transformer models for time series prediction that incorporate exponential smoothing.

We monitor the price and demand for electricity using weather and socioeconomical features, by first detecting anomalies then forecasting.

Predict short-term loads accurately by leveraging federated learning & incremental changes in weather and historical load pattern over time

Explainable probabilistic model which uses smooth periodic copula; handles temporal dependencies; can do anomaly detection and forecasting.

An accurate depiction of the uncertainty of day-ahead net load by leveraging advanced statistical techniques & machine learning algorithms

Vizion Forecaster uses an advanced Machine Learning model to forecast the net load.

Data Cleaning Quantile Multiple Linear Regression Deep NN Re-train over-time

A deep learning-based method for accurately forecasting substation net load, using Long Short-Term Memory (LSTM).

Forecasts will be created by combining regression and non-regression models to synthesize all predictability in the data.

The team will develop data-driven models to predict. We preprocess the data, do feature engineering, train models, and ensemble all models.

We propose an architecture for Net Load Forecasting leveraging deep neural network technology and reinforcement learning techniques.