Outlier Impact Characterization for Time Series Data

Authors: Jianbo Li, Lecheng Zheng, Yada Zhu, Jingrui He11595-11603

AAAI 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The empirical results on multiple real data sets demonstrate the effectivenss of the proposed SIF metric.In this section, we quantitatively evaluate the influence functional (IF) and the SIF with respect to the impact of the contaminating process on the parameter estimation and future predictions.
Researcher Affiliation Collaboration Jianbo Li, 1 Lecheng Zheng, 2 Yada Zhu, 3 Jingrui He 2 1 Three Bridges Capital, 2 University of Illinois at Urbana-Champaign, 3 IBM Research jianboliru@gmail.com, {lecheng4, jingrui}@illinois.edu, yzhu@ibm.us.com
Pseudocode No No structured pseudocode or algorithm blocks were found in the paper.
Open Source Code No One can verify all the experimental results using our code, to be shared upon paper acceptance.
Open Datasets Yes For the semi-synthetic data, a clean time series, real 35, is randomly selected from the Yahoo! Webscope and contaminated in the same way as the synthetic data. The original ECG data is obtained from https://physionet.org/content/chfdb/1.0.0/.
Dataset Splits No The paper mentions splitting data into train and test sets (60:40) but does not provide specific details for a validation set split.
Hardware Specification Yes All the experiments in this work are done on a Macbook Pro with Intel(R) Core(TM) i7 CPU and 16 GB memory.
Software Dependencies Yes The code is written under Python 3.6 with Tensor Flow 1.12.
Experiment Setup Yes We set the coefficient of the core process as 0.7 and γ [0, 0.01].We randomly select the coefficients for the ARMA(2, 2) model under the constraint of the stationary triangle (J urgen Franke 2015), and choose the LSTM and RNN models with two layers and 256 hidden states.To this end, we determine the optimal ARMA coefficients by maximizing the SIF2 under the constraint of the stationary triangle. We solve the constrained optimization problem using SLSQP (Kraft 1988), a standard package available in Python.