GeomComb (version 1.0)

comb_LAD: Least Absolute Deviation Forecast Combination

Description

Computes forecast combination weights using least absolute deviation (LAD) regression.

Usage

comb_LAD(x)

Arguments

x

An object of class 'foreccomb'. Contains training set (actual values + matrix of model forecasts) and optionally a test set.

Value

Returns an object of class foreccomb_res with the following components: with the following components:

Details

The function is a wrapper around the least absolute deviation (LAD) forecast combination implementation of the ForecastCombinations package.

The defining property of comb_LAD is that it does not minimize the squared error loss like comb_OLS and comb_CLS, but the absolute values of the errors. This makes the method more robust to outliers -- comb_LAD tends to penalize models, which have high errors for some observations, less harshly than the least squares methods would.

Optimal forecast combinations under general loss functions are discussed by Elliott and Timmermann (2004). The LAD method is described in more detail, and used in an empirical context, by Nowotarksi et al. (2014).

The results are stored in an object of class 'foreccomb_res', for which separate plot and summary functions are provided.

References

Elliott, G., and Timmermann, A. (2004). Optimal Forecast Combinations Under General Loss Functions and Forecast Error Distributions. Journal of Econometrics, 122(1), 47--79.

Nowotarski, J., Raviv, E., Tr\"uck, S., and Weron, R. (2014). An Empirical Comparison of Alternative Schemes for Combining Electricity Spot Price Forecasts. Energy Economics, 46, 395--412.

See Also

Forecast_comb, foreccomb, plot.foreccomb_res, summary.foreccomb_res, accuracy

Examples

obs <- rnorm(100)
preds <- matrix(rnorm(1000, 1), 100, 10)
train_o<-obs[1:80]
train_p<-preds[1:80,]
test_o<-obs[81:100]
test_p<-preds[81:100,]

data<-foreccomb(train_o, train_p, test_o, test_p)
comb_LAD(data)