A Robust Functional EM Algorithm for Incomplete Panel Count Data
Authors: Alexander Moreno, Zhenke Wu, Jamie Roslyn Yap, Cho Lam, David Wetter, Inbal Nahum-Shani, Walter Dempsey, James M. Rehg
NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We illustrate the utility of the proposed algorithm through numerical experiments and an analysis of smoking cessation data. We analyze two datasets empirically. |
| Researcher Affiliation | Academia | Alexander Moreno Georgia Institute of Technology alexander.f.moreno@gatech.edu Zhenke Wu University of Michigan zhenkewu@umich.edu Jamie Yap University of Michigan jamieyap@umich.edu Cho Lam University of Utah cho.lam@hci.utah.edu David W. Wetter University of Utah david.wetter@hci.utah.edu Inbal Nahum-Shani University of Michigan inbal@umich.edu Walter Dempsey University of Michigan wdem@umich.edu James M. Rehg Georgia Institute of Technology rehg@gatech.edu |
| Pseudocode | Yes | Algorithm 1 Sample-based EM Algorithm for Panel Count Data with Missing Counts. |
| Open Source Code | No | The paper does not provide an explicit statement or a link to its source code for the described methodology. It mentions that [30, 7] provide software implementations of AEE, but this refers to third-party code, not the authors' implementation. |
| Open Datasets | Yes | We first look at a bladder tumor dataset where there are no missing intervals... [5] BYAR, D. The veterans administration study of chemoprophylaxis for recurrent stage i bladder tumours: comparisons of placebo, pyridoxine and topical thiotepa. In Bladder tumors and other topics in urological oncology. Springer, 1980, pp. 363 370. |
| Dataset Splits | No | The paper mentions analyzing two datasets but does not specify any training, validation, or test dataset splits (e.g., percentages or sample counts). It refers to bootstrapping for analysis rather than data partitioning for model validation. |
| Hardware Specification | No | The paper does not provide any specific details about the hardware used to run the experiments, such as CPU or GPU models, memory, or cloud computing specifications. |
| Software Dependencies | No | The paper mentions using a 'general likelihood-based augmented estimating equation (AEE) method [29]' but does not specify version numbers for any software dependencies, libraries, or programming languages used in their implementation or experiments. |
| Experiment Setup | Yes | In the M-step, we can choose any reasonable mean function estimator. In our experiments, we use a general likelihood-based augmented estimating equation (AEE) method [29], which uses monotone step functions when obtaining the mean function estimate with complete data. We artificially delete intervals completely at random with probability 0.2. We then initialize Λ(0) by replacing the missing data with Poisson(1) random variables and fitting AEE. We bootstrap 1,000 times, and plot the sample mean of our learned mean functions under complete data. We use the counts to initialize our model. |