Online Time Series Prediction with Missing Data

Authors: Oren Anava, Elad Hazan, Assaf Zeevi

ICML 2015 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We show that our algorithm s performance asymptotically approaches the performance of the best AR predictor in hindsight, and corroborate the theoretic results with an empirical study on synthetic and real-world data. 4. Illustrative Examples The following experiments demonstrate the effectiveness of the proposed algorithm under various synthetic settings.
Researcher Affiliation Academia Oren Anava OANAVA@TX.TECHNION.AC.IL Technion, Haifa, Israel Elad Hazan EHAZAN@CS.PRINCETON.EDU Princeton University, NY, USA Assaf Zeevi ASSAF@GSB.COLUMBIA.EDU Columbia University, NY, USA
Pseudocode Yes Algorithm 1 LAZY OGD (on ℓ2-ball with radius D), Algorithm 2, Algorithm 3 Efficient Implementation of Algorithm 2, Algorithm 4 OGDIMPUTE
Open Source Code No The paper does not contain any explicit statements about releasing source code or links to a code repository.
Open Datasets No The paper uses 'synthetic data' for its experiments, as described in Section 4.2. While it mentions 'real-world data' in the abstract, no specific public dataset name, link, or citation is provided for either the synthetic or real-world data used.
Dataset Splits No The paper conducts experiments but does not explicitly provide information on dataset splits such as train/validation/test percentages, absolute sample counts for splits, or cross-validation setup.
Hardware Specification No The paper does not provide specific hardware details such as GPU or CPU models, memory specifications, or cloud instance types used for running the experiments.
Software Dependencies No The paper does not specify any software dependencies with version numbers (e.g., programming languages or libraries with their exact versions).
Experiment Setup No The paper states 'For our algorithm, we used d = 3p in all considered settings' and that results are 'averaged over 50 runs'. However, it does not provide other specific experimental setup details such as learning rates, batch sizes, optimizer settings, or number of epochs for the models used.