Please enable JavaScript.
Coggle requires JavaScript to display documents.
Performance Evaluation (Prediction Accuracy Measures (forecast error…
Performance Evaluation
Overfitting
Fit systematic component of the data
Fit the noise
Data Partitioning
Cross-sectional data
Training set
build models
Validation set
assess performance of each model
Test set
assess performance of the chosen model with new data
Time series
Training period
train methods on the earlier period
Validation period
assess predictive performance on the later period
Roll-forward validation
Naive Forecasts
the most recent value of the series
seasonal naive forecast: the value from the most recent identical season
Prediction Accuracy Measures
forecast error (residual)
MAE or MAD (mean absolute error/deviation)
Average error
MAPE (mean absolute percentage error)
RMSE (root mean squared error)