2019
DOI: 10.1111/rssb.12317
|View full text |Cite
|
Sign up to set email alerts
|

MALMEM: Model Averaging in Linear Measurement Error Models

Abstract: Summary We develop model averaging estimation in the linear regression model where some covariates are subject to measurement error. The absence of the true covariates in this framework makes the calculation of the standard residual‐based loss function impossible. We take advantage of the explicit form of the parameter estimators and construct a weight choice criterion. It is asymptotically equivalent to the unknown model average estimator minimizing the loss function. When the true model is not included in th… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1

Citation Types

0
3
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
6

Relationship

0
6

Authors

Journals

citations
Cited by 13 publications
(3 citation statements)
references
References 42 publications
0
3
0
Order By: Relevance
“…To further improve the predictive power of the transformation methods, we treat DET, DIT, and PCT as three weak learners and follow the well‐known ensemble learning approach, model averaging (Hansen, 2007; Zhang et al ., 2019), to obtain enhanced predictions. As shown in Figure 2, ensemble learning uses the base models' outputs as the input features and the predicted variable remains the same as in the original data set.…”
Section: Predictive Analytics: Nonstationary Transformation Methodsmentioning
confidence: 99%
See 1 more Smart Citation
“…To further improve the predictive power of the transformation methods, we treat DET, DIT, and PCT as three weak learners and follow the well‐known ensemble learning approach, model averaging (Hansen, 2007; Zhang et al ., 2019), to obtain enhanced predictions. As shown in Figure 2, ensemble learning uses the base models' outputs as the input features and the predicted variable remains the same as in the original data set.…”
Section: Predictive Analytics: Nonstationary Transformation Methodsmentioning
confidence: 99%
“…Moreover, the validation analysis shows that the three methods also provide satisfying predictive performance when compared with the benchmark approaches. To further improve the fitting effect of the transformed methods, we treat DET, DIT, and PCT as three weak learners and follow the well‐known ensemble learning approach, model averaging (Hansen, 2007; Zhang et al ., 2019), to obtain enhanced predictions for the nonstationary demand. Specifically, the model averaging approach assigns a positive weight to every transformation method (the sum of three weights is 1) and outputs the weighted sum of their predictions as the enhanced predictions.…”
Section: Introductionmentioning
confidence: 99%
“…It can be considered as an extreme of imputation in the missing value problem, where all cases are missing for some predictors (see also Boone et al ., 2011). The measurement error model also has a similar structure, in that the true value is unobserved (see also Doppelhofer et al ., 2016; Zhang et al ., 2019).…”
Section: Economic Variable Selectionmentioning
confidence: 99%