Causal Inference in Threshold Regression and the Neural Network Extension (TRNN)
The first-hitting-time based model conceptualizes a random process for subjects’ latent health status. The time-to-event outcome is modeled as the first hitting time of the random process to a pre-specified threshold. Threshold regression with linear predictors has numerous benefits in causal surviv...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Stats |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2571-905X/6/2/36 |
| _version_ | 1797592573852778496 |
|---|---|
| author | Yiming Chen Paul J. Smith Mei-Ling Ting Lee |
| author_facet | Yiming Chen Paul J. Smith Mei-Ling Ting Lee |
| author_sort | Yiming Chen |
| collection | DOAJ |
| description | The first-hitting-time based model conceptualizes a random process for subjects’ latent health status. The time-to-event outcome is modeled as the first hitting time of the random process to a pre-specified threshold. Threshold regression with linear predictors has numerous benefits in causal survival analysis, such as the estimators’ collapsibility. We propose a neural network extension of the first-hitting-time based threshold regression model. With the flexibility of neural networks, the extended threshold regression model can efficiently capture complex relationships among predictors and underlying health processes while providing clinically meaningful interpretations, and also tackle the challenge of high-dimensional inputs. The proposed neural network extended threshold regression model can further be applied in causal survival analysis, such as performing as the Q-model in G-computation. More efficient causal estimations are expected given the algorithm’s robustness. Simulations were conducted to validate estimator collapsibility and threshold regression G-computation. The performance of the neural network extended threshold regression model is also illustrated by using simulated and real high-dimensional data from an observational study. |
| first_indexed | 2024-03-11T01:56:02Z |
| format | Article |
| id | doaj.art-7a05d1c7fa234ece853e607cf84997a9 |
| institution | Directory Open Access Journal |
| issn | 2571-905X |
| language | English |
| last_indexed | 2024-03-11T01:56:02Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Stats |
| spelling | doaj.art-7a05d1c7fa234ece853e607cf84997a92023-11-18T12:39:16ZengMDPI AGStats2571-905X2023-04-016255257510.3390/stats6020036Causal Inference in Threshold Regression and the Neural Network Extension (TRNN)Yiming Chen0Paul J. Smith1Mei-Ling Ting Lee2Department of Epidemiology and Biostatistics, University of Maryland, College Park, MD 20742, USADepartment of Mathematics, University of Maryland, College Park, MD 20742, USADepartment of Epidemiology and Biostatistics, University of Maryland, College Park, MD 20742, USAThe first-hitting-time based model conceptualizes a random process for subjects’ latent health status. The time-to-event outcome is modeled as the first hitting time of the random process to a pre-specified threshold. Threshold regression with linear predictors has numerous benefits in causal survival analysis, such as the estimators’ collapsibility. We propose a neural network extension of the first-hitting-time based threshold regression model. With the flexibility of neural networks, the extended threshold regression model can efficiently capture complex relationships among predictors and underlying health processes while providing clinically meaningful interpretations, and also tackle the challenge of high-dimensional inputs. The proposed neural network extended threshold regression model can further be applied in causal survival analysis, such as performing as the Q-model in G-computation. More efficient causal estimations are expected given the algorithm’s robustness. Simulations were conducted to validate estimator collapsibility and threshold regression G-computation. The performance of the neural network extended threshold regression model is also illustrated by using simulated and real high-dimensional data from an observational study.https://www.mdpi.com/2571-905X/6/2/36first-hitting-timeG-computationmachine learninghigh-dimensional data |
| spellingShingle | Yiming Chen Paul J. Smith Mei-Ling Ting Lee Causal Inference in Threshold Regression and the Neural Network Extension (TRNN) Stats first-hitting-time G-computation machine learning high-dimensional data |
| title | Causal Inference in Threshold Regression and the Neural Network Extension (TRNN) |
| title_full | Causal Inference in Threshold Regression and the Neural Network Extension (TRNN) |
| title_fullStr | Causal Inference in Threshold Regression and the Neural Network Extension (TRNN) |
| title_full_unstemmed | Causal Inference in Threshold Regression and the Neural Network Extension (TRNN) |
| title_short | Causal Inference in Threshold Regression and the Neural Network Extension (TRNN) |
| title_sort | causal inference in threshold regression and the neural network extension trnn |
| topic | first-hitting-time G-computation machine learning high-dimensional data |
| url | https://www.mdpi.com/2571-905X/6/2/36 |
| work_keys_str_mv | AT yimingchen causalinferenceinthresholdregressionandtheneuralnetworkextensiontrnn AT pauljsmith causalinferenceinthresholdregressionandtheneuralnetworkextensiontrnn AT meilingtinglee causalinferenceinthresholdregressionandtheneuralnetworkextensiontrnn |