Abstract:In the usual statistical inference problem, we estimate an unknown parameter of a statistical model using the information in the random sample. A priori information about the parameter is also known in several real-life situations. One such information is order restriction between the parameters. This prior formation improves the estimation quality. In this paper, we deal with the component-wise estimation of location parameters of two exponential distributions studied with ordered scale parameters under a bowl-shaped affine invariant loss function and generalized Pitman closeness criterion. We have shown that several benchmark estimators, such as maximum likelihood estimators (MLE), uniformly minimum variance unbiased estimators (UMVUE), and best affine equivariant estimators (BAEE), are inadmissible. We have given sufficient conditions under which the dominating estimators are derived. Under the generalized Pitman closeness criterion, a Stein-type improved estimator is proposed. As an application, we have considered special sampling schemes such as type-II censoring, progressive type-II censoring, and record values. Finally, we perform a simulation study to compare the risk performance of the improved estimators
What problem does this paper attempt to address?
### Problems the Paper Attempts to Solve
This paper aims to address the problem of estimating location parameters in two exponential distributions with ordered scale parameters. Specifically, the paper investigates how to estimate the location parameters of two exponential distributions under the bowl-shaped affine invariant loss function and the generalized Pitman nearness criterion.
### Background and Motivation
In statistical inference problems, it is often necessary to estimate unknown parameters using information from random samples. In many practical applications, prior information about the parameters is also known, one common type of prior information being the order restrictions between parameters. This prior information can improve the quality of the estimates. For example, in reliability theory, exponential distributions are commonly used to model the lifetimes of experimental units, where the location parameter represents the guaranteed time of a component, and the reciprocal of the scale parameter represents the instantaneous failure rate or hazard rate. Therefore, the order relationship between scale parameters naturally exists in life testing experiments.
### Research Content
The paper mainly investigates the following points:
1. **Estimation Methods**:
- Discusses the inadmissibility of Maximum Likelihood Estimation (MLE), Uniformly Minimum Variance Unbiased Estimation (UMVUE), and Best Affine Equivariant Estimation (BAEE).
- Proposes improved estimators and provides sufficient conditions for these estimators under specific conditions.
2. **Loss Functions**:
- Uses bowl-shaped affine invariant loss functions and the generalized Pitman nearness criterion to evaluate the performance of the estimators.
3. **Special Sampling Schemes**:
- Considers special sampling schemes such as Type II censoring, progressive Type II censoring, and record values.
4. **Simulation Studies**:
- Conducts simulation studies to compare the risk performance of the improved estimators.
### Main Contributions
- **Theoretical Contributions**: Demonstrates the inadmissibility of several benchmark estimators (such as MLE, UMVUE, and BAEE) and proposes improved estimators.
- **Applied Contributions**: Shows the effectiveness of the improved estimators in practical applications by considering special sampling schemes.
- **Methodological Contributions**: Proposes Stein-type improved estimators under the generalized Pitman nearness criterion and discusses their performance under different loss functions.
### Conclusion
By introducing new estimation methods and loss functions, this paper improves the quality of estimating location parameters of exponential distributions under ordered scale parameters. These methods have significant applications in fields such as reliability engineering, clinical trials, and biometrics.