ProLoaF: v0.2.2

A Probabilistic Load Forecasting Project

v0.1.1

• Updated User Guide
• Added Functionalities to explain importance of features in training (Draft version)
• Fixed smaller issues, such as links and consistent naming
• Updated Benchmark Models
• Added new Plot Functionalities, such as error distribution plots
• Updated Plotting and Evaluation to compare multiple models with each other
• Updates to Code Documentation now available at: https://sogno-platform.github.io/proloaf/
• New example data added

v0.2.2

• fixed bugs validation_ds, quantile prediction failing, hparams updating
• updated interpreter functionalities
• included structure for transformer, informer models
• added smoothed pinball loss
• updated example notebooks (includes more benchmarks)
• created torch dataloader, tensorloader, which assembles all data prep steps and postpones jobs to torch functions rather than requiring numpy or pandas routines up front