Multivariate Time Series Imputation with Generative Adversarial Networks
Authors: Yonghong Luo, Xiangrui Cai, Ying ZHANG, Jun Xu, Yuan xiaojie
NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments on two multivariate time series datasets show that the proposed model outperformed the baselines in terms of accuracy of imputation. |
| Researcher Affiliation | Academia | Yonghong Luo College of Computer Science Nankai University Tianjin, China luoyonghong@dbis.nankai.edu.cn Xiangrui Cai College of Computer Science Nankai University Tianjin, China caixiangrui@dbis.nankai.edu.cn Ying Zhang College of Computer Science Nankai University Tianjin, China yingzhang@nankai.edu.cn Jun Xu School of Information Renmin University of China Beijing, China junxu@ruc.edu.cn Xiaojie Yuan College of Computer Science Nankai University Tianjin, China yuanxj@nankai.edu.cn |
| Pseudocode | No | The paper does not contain any pseudocode or clearly labeled algorithm blocks. |
| Open Source Code | No | The paper does not provide any concrete access information (e.g., repository link or explicit statement) for open-source code for the methodology it describes. |
| Open Datasets | Yes | Physionet Challenge 2012 dataset (Physio Net). The Physionet dataset is a public electronic medical record dataset that comes from the Physio Net Challenge 2012 [42]. KDD CUP 2018 Dataset (KDD). The KDD CUP 2018 dataset is a public air quality dataset that comes from the KDD CUP Challenge 2018 [11]. |
| Dataset Splits | No | The paper does not explicitly provide specific training/test/validation dataset splits (e.g., percentages, sample counts, or references to predefined splits). |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts) used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiment. |
| Experiment Setup | Yes | For the Physio Net dataset, the input dimension is 41 (we use all the variables of the Physio Net dataset), batch size is 128, the hidden units number in GRUI of G and D is 64 and the dimension of random noise is also 64. For the KDD dataset, the input dimension is 132 (11 observatories 12 variables), the batch size is 16, the number of hidden units in GRUI of G and D is 64 and the dimension of z is 256. The hyper-parameters of our method are: the train epochs is 30, pretrain epochs is 5, learning rate is 0.001, λ is 0.15 and the number of optimization iterations of the imputation loss is 400. The hyper-parameters of our method are: the train epochs is 25, pretrain epochs is 20, learning rate is 0.002, λ is 0.0 and the number of optimization iterations of the imputation loss is 800. |