Questions tagged [diebold-mariano-test]

A test for superior predictive accuracy of one forecast over another.

Filter by
Sorted by
Tagged with
0 votes
0 answers
22 views

Interpreting Diebold-Mariano results

I preformed the DM test of two series on Python and got this result: DM=-0.8149537766612477, p_value=0.4170549576990228) How should I make the interpretation of ...
user avatar
1 vote
1 answer
24 views

Is ARIMA-GARCH nested within ARIMA?

I wanted to compare ARIMA(1,1,1)-GARCH(1,1) and ARIMA(1,1,1) model forecasts with a Diebold-Mariano test, but I know that it cannot be used for nested models. Is ARIMA-GARCH technically nested within ...
user avatar
  • 25
0 votes
0 answers
18 views

Statistical test for forecast performance over multiple runs

Lets say I have a time series, create a training and test set, and I want to compare the predictive accuracy of two models, by measuring e.g. the mean absolute error (MAE) over the test set. I know ...
user avatar
1 vote
0 answers
22 views

How to compare different models in their ability to forecast the value-at-risk with Diebold-Mariano test?

I made value-at-risk forecasts for different models for the 95, 97.5 and 99%. I also made a dummy which equals 1 if the true return was below the value-at-risk, 0 otherwise. How can I compare those ...
user avatar
  • 41
0 votes
0 answers
12 views

Statistical Tests for Models Performance using Multiple Simulated Data sets

In my project, I want to test whether Model 1 (M1) has equal or better performance than Model 2 (M2) using simulated datasets. Given the natural of the problem, I use the Diebold-Mariano test. The ...
user avatar
  • 1
0 votes
0 answers
51 views

Diebold-Mariano test for comparing multivariate models

I want to know if there exists a Diebold-Mariano test version for comparing the accuracy of predictions between multivariate statistics models. For example, compare a VAR model and multivariate GARCH ...
user avatar
  • 211
1 vote
1 answer
199 views

Meaning of Diebold-Mariano (DM) test for other accuracy measures (MDA, $R^2$...)

I am trying to compare the accuracies of two time series forecasts. I read about the Diebold-Mariano (DM) Test, which tests the null hypothesis of $E[d_t]=0$, where $d_t=g(e_{it})-g(e_{jt})$ is the ...
user avatar
0 votes
1 answer
30 views

Vector valued time series forecast significance testing?

I have two vector valued time series forecasts, and the components of the vector can be correlated. How do i significance test (ex. diebold mariano) for which is the better forecaster? I found two ...
user avatar
1 vote
1 answer
252 views

time series Diebold-Mariano Test n-ahead forecast in R

I have a question about "h" in dm.test & DM.test from package forecast & multDM, assume I set a ARIMA model to forecast n-ahead = 20 (using dynamic regression not one step forecast), ...
user avatar
  • 69
1 vote
0 answers
64 views

CLT in Diebold & Mariano (1995)

The Diebold-Mariano (DM) statistic is derived as follows: Assuming the loss-differential between the two models $d_t$ is covariance stationary: $$\begin{cases} \mathbb{E}[d_t] = \mu> 0 & \...
user avatar
8 votes
2 answers
472 views

How can I determine that a forecast is significantly more accurate than another one? (time series)

Reproducible Example Look at this reproducible example: I have a time series that I want to forecast. For the sake of reproducibility, I'll just take AirPassanger. ...
user avatar
  • 183
1 vote
1 answer
152 views

Diebold Mariano test for non time-series data

Can the Diebold Mariano test be used to compare the MSE of two models (trained on the same data) for non time series data?
user avatar
  • 11
4 votes
2 answers
421 views

How to test superior predictive ability over multiple time series?

Suppose you have two models, model A and model B, and let these models forecast 10 time series over a horizon of 12 periods. That is, suppose the time series contain monthly data and your forecasting ...
user avatar
  • 282
1 vote
0 answers
278 views

Diebold-Mariano test for predictive accuracy with R

The following example code is from the documentation of the command dm.test from R. ...
user avatar
0 votes
1 answer
225 views

Specifying h in dm.test function in R

I have a question about the use of the dm.test function from the forecast package in R. Starting from ...
user avatar
1 vote
1 answer
328 views

Should I use the Diebold-Mariano test year-by-year or on the overall forecast?

I have built two models, one ARIMAX and one VAR, to compare against a baseline ARIMA model to predict a weekly economic time series of interest. I am primarily comparing the accuracy of my models ...
user avatar
4 votes
1 answer
362 views

t-test for time series (Diebold Mariano test?)

Currently I am working on the following problem: I want to compare the means of two different time series ts1 and ts2. As the samples are not iid and t-test can't be applied, I thought to use the ...
user avatar
  • 41
5 votes
1 answer
690 views

Alternatives for Diebold-Mariano test when comparing the best forecast among many against a benchmark

Suppose I encounter a new forecasting method and I wish to evaluate it against a benchmark. I can obtain forecasts from the two methods and compare them to actual realizations and thus obtain the ...
user avatar
1 vote
0 answers
129 views

Is it possible to bootstrap a Diebold-Mariano test?

I am currently working on a small project where I want to use a (two-sided) 1-step ahead ($h=1$) Diebold Mariano test to compare forecast losses for different realized measures calculated on time-...
user avatar
0 votes
0 answers
145 views

Diebold Mariano test Nested Models

I have computed forecasts with 4 different methods, namely OLS, Elastic Net, Cubic splines in combination with Lasso, and Neural Network. All models use the same set of base variables, except cubic ...
user avatar
  • 193
2 votes
1 answer
270 views

Can I give continuous rank probability score (CRPS) to Diebold-Mariano (DM) test?

I would like to use DM test for probabilistic forecasting case. My initial thinking was to give CRPS of two forecasting methods instead of raw forecast errors, where CRPS is calculated using ...
user avatar
  • 68
0 votes
0 answers
303 views

Autocorrelation in loss function, want to perform DM test. How many lags to use? (R)

I have two sets of forecasting errors, and want to perform a DM test. Both forecasts are a fixed size moving window, and are 1 day ahead forecasts. The first step of performing the DM test is to ...
user avatar
  • 83
5 votes
1 answer
3k views

How to test if two RMSE are significantly different?

Say I have two models for a regression task and from each model I get a RMSE. One RMSE is smaller than the other, however I wish to test if the difference is statistically significant in order to be ...
user avatar
2 votes
0 answers
185 views

Can i use Diebold Mariano test for comparison of 2 models across multiple time series?

I have 2 models (for simplicity, let's call them AR(1) and MA(1)) each making 1 day ahead forecasts of time series. If I had only 1 time series I would just use ...
user avatar
  • 399
3 votes
1 answer
310 views

Diebold-Mariano for in-sample model comparisons?

Suppose I have two in-sample forecasts from two different non-nested models. I want to check which one produces the best forecasts. A common way is Diebold-Mariano, GiacominiWhite, ENC-T test. However,...
user avatar
0 votes
1 answer
219 views

Comparing Forecast Errors

I want to compare my VAR-forecast to an ARMA-benchmark. The VAR forecast has a smaller MPSE (mean squared prediction error) and I want to check for significance. For this purpose I applied the Diebold-...
user avatar
  • 1
1 vote
1 answer
156 views

Diebold-Mariano with multiple predictions over time

I've been trying to use a DM test on my predictive model. I have two different models. Each model is calibrated by using a MSE loss function to estimate the price of 374 different assets each day. I ...
user avatar
3 votes
1 answer
2k views

Diebold-Mariano in the context of volatility forecasting: What is the ultimate aim of this test?

Perhaps I'm missing a simple conceptual point here. But do the error statistics (RMSE, MAE) not tell which is the best forecast by presenting the lowest figures between the forecast and the actual ...
user avatar
  • 109
2 votes
1 answer
711 views

Tests of Forecast Accuracy for Nested Models

Can anyone explain why "classic" tests of forecast accuracy, (i.e. Diebold-Mariano test, Meese-Rogoff test and Morgan-Granger-Newbold test) are not suited for nested models? I could not find a good ...
user avatar
  • 21
1 vote
1 answer
2k views

Compute the p-value of Diebold-Mariano Test

I am using Diebold-Mariano Test for testing the equal predictive accuracy of two models. I use the code written by Semin Ibisevic (2011) in MATLAB to compute it ...
user avatar
  • 13
1 vote
0 answers
330 views

Diebold-Mariano test in Stata, autocovariance denominator doubt

I'm trying to reproduce the Diebold&Mariano test available in Stata13 and I think there is an error in the suggested implementation. At the bottom of the file you have: ...
user avatar
4 votes
1 answer
1k views

Diebold-Mariano with RMSE as loss function?

When applying the Diebold-Mariano test to test for predictive accuracy we need to specify a loss differential. For instance the loss differential $d$ in terms of the mean absolute error (MAE) is $d = ...
user avatar
1 vote
1 answer
2k views

Diebold-Mariano, but how to say which one is more accurate?

I know that the Diebold-Mariano test is simply testing whether two forecasts are likely to forecast with the same accuracy (null hypothesis) or not (reject the null) based on some loss function. But ...
user avatar
1 vote
0 answers
39 views

Test if one forecast outperforms all other

I have 4 different models to forecast some quantity of interest y With the Diebold-Mariano Test I can test if one model is predicting more accurate than another, which is cool. So I can make some 4x4 ...
user avatar
  • 247
2 votes
1 answer
3k views

Diebold-Mariano test for multiple prediction horizons

I am trying to compare two forecasts using the Mariano Diebold test in R. Both forecasts are for 150 days ahead; that is, on day $t$ I forecast $t+1, t+2, \dotsc, t+150$. I deduced from this post ...
user avatar
  • 33
1 vote
1 answer
2k views

Diebold-Mariano test in case of nested models (Clark & McCracken, 2001)

I have become aware of Clark & McCracken (2001) showing that the asymptotics of the Diebold-Mariano test will potentially collapse when comparing forecast accuracy of nested models (such as GARCH /...
user avatar
0 votes
0 answers
686 views

Diebold-Mariano Test in R - how to rationalize p-value?

Consider 3 models (model1, model2, model3) with the following set of prediction errors: ...
user avatar
  • 574
2 votes
1 answer
1k views

Diebold-Mariano test: robust standard errors for 1-day ahead forecasts?

Diebold & Mariano suggest using Newey-West standard errors to correct for autocorrelation and heteroscedasticity in the error terms when comparing forecast accuracy. However, if I understand ...
user avatar
1 vote
1 answer
5k views

Diebold Mariano test (in R)

As I asked in my answer to this question: does anyone know if the DM test (in R in this case) is supposed to be made with h=h-1? If not, am I supposed to make several prediction sets (with h ...
user avatar
  • 488
0 votes
1 answer
551 views

Diebold - Mariano test for volatility forecasts problem

I am using packages {rugarch} for forecasting and {forecast} for Diebold - Mariano test. As a first step, I am specifying the first AR-GARCH model for financial time series (AAPL Nasdaq) using ...
user avatar
  • 1
3 votes
1 answer
419 views

Difference between comparing forecasts and models

I started out looking for a way to test the difference between MSPE between two models (Question here), when (thanks to @Richard Hardy) I ended up reading a paper of Diebold regarding the Diebold-...
user avatar
  • 1,434
4 votes
1 answer
1k views

ANOVA / t-test to compare the errors of different models

I have two forecasting models, moving average and single exponential smoothing. The values of Mean Absolute Percentage Error (MAPE) is 5.2%, 5.8%. Since the difference of MAPE between the models are ...
user avatar
  • 205
1 vote
1 answer
970 views

On a problem with the implementation of the test of Diebold and Mariano for equal predictive accuracy

To recall how the test of Diebold and Mariano is set up I utilize the parsimonious words of Diebold himself: Now, I have been trying to perform this test on the respective vector of errors of two ...
user avatar
  • 1,101
2 votes
1 answer
268 views

Using Diebold-Mariano test to compare predictive errors in non-time-series?

I understand that the DM test is established for time series data, but could I still apply the test for non-time-series data? Could I simply replace the autocorvaiance part of the test statistics with ...
user avatar
  • 125
7 votes
1 answer
9k views

Diebold-Mariano test for predictive accuracy

I am using the Diebold-Mariano test in the forecast package of R to test the predictive accuracy. In particular, I want to ...
user avatar
2 votes
3 answers
3k views

Understanding forecast horizon for Diebold-Mariano tests

I have a problem understanding the parameter horizon of the function dm.test {forecast} in my particular setting. Background: My goal is to forecast energy consumption for individual households. The ...
user avatar