GroupSegment-SHAP: Shapley Value Explanations with Group-Segment Players for Multivariate Time Series

Jinwoong Kim, Sangjin Park

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 16 functions out of this paper's own repositories and ran 11 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

Repositories linked to this paper

• jirungx/GroupSegment-SHAP

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Multivariate time-series models achieve strong predictive performance in healthcare, industry, energy, and finance, but how they combine cross-variable interactions with temporal dynamics remains unclear. SHapley Additive exPlanations (SHAP) has been widely used for interpretation. However, existing time-series variants typically treat the feature and time axes independently, fragmenting structural signals formed jointly by multiple variables over specific intervals. We propose GroupSegment-SHAP (GS-SHAP), which constructs explanatory units as group-segment players based on cross-variable dependence and distribution shifts over time, and then quantifies each unit's contribution via Shapley attribution. We evaluated GS-SHAP across four real-world domains: human activity recognition, power-system forecasting, medical signal analysis, and financial time series, and compared it with KernelSHAP, TimeSHAP, SequenceSHAP, WindowSHAP, and TSHAP. GS-SHAP consistently achieves the highest deletion-based faithfulness (ΔAUC) across all four benchmark datasets. In synthetic evaluations with controlled ground-truth structures, GS-SHAP achieves 24.4% higher IoU-based recovery of multivariate-temporal patterns than the strongest baseline. In addition, GS-SHAP runs on average 3.38× faster than Time-SHAP across approximation budgets in power-system forecasting, showing that it can simultaneously achieve high explanatory faithfulness and computational efficiency. In a financial case study, GS-SHAP identifies interpretable multivariate-temporal interactions among key market variables across market regimes, highlighting its potential utility for risk-aware investment analysis. Code is available at https://github.com/jirungx/GroupSegment-SHAP. • Computing methodologies → Machine learning; Neural networks; • Mathematics of computing → Time series analysis; • Information systems → Data mining.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2601.06114")
get_code_for_paper("2601.06114")
have("2601.06114")

Connect an agent — have() is free.