This is a case study of forecasting short-term electricity loads for the Australian market.
%% Neural Network Load Forecasting
% This short script builds a Neural Network regression model for
% predicting day-ahead load from a predictor matrix consisting of
% temperature, date/time and lagged load data.
%% Initialize and Train Network
% Initialize a default network of two layers with 20 neurons. Use the "mean
% absolute error" (MAE) performance metric. Then, train the network with
% the default Levenburg-Marquardt algorithm. For
net = newfit(trainX', trainY', 20);
net.performFcn = 'mae';
%% Train Network
% Train the network using the default Levenburg-Marquardt algorithm
net = train(net, trainX', trainY');
%% Predict with Network
% Perform a forecast on the independent testing dataset.
forecastLoad = sim(net, testX')';