static
TemporalScope/src/temporalscope/partition/single_target/static/init.py.
This module defines the namespace for static partitioning algorithms for single-target workflows in TemporalScope. Static partitioning refers to user-defined, fixed, and deterministic strategies, contrasting with dynamic (algorithmic or data-driven) methods. These methods are designed to work seamlessly with TimeFrame objects and support backend-agnostic operations via the Narwhals ecosystem.
Static Partitioning:
Static partitioning encompasses human-defined techniques, where partitions are specified based on domain knowledge or explicit rules. These methods are suited for: - Sliding Window Partitioning: Dividing data into fixed-size, non-overlapping or overlapping partitions. - Expanding Window Partitioning: Creating sequentially increasing partitions for cumulative analysis.
Single-Target Workflows:
This namespace supports DataFrame-centric workflows for scalar target models. Multi-target workflows, as described in the TemporalPartitionerProtocol, will require further extensions to ensure compatibility with TensorFlow/PyTorch datasets and sequence-target workflows.
Extensibility:
TemporalScope provides the foundation for users to implement or customize their own static partitioning strategies. This flexibility ensures the framework remains adaptable to diverse requirements and emerging techniques in time-series analysis.
Notes
Users are encouraged to leverage the TemporalPartitionerProtocol for building custom static partitioning workflows and refer to the foundational literature on partitioning techniques for guidance.
See Also
- Shah, A., DePavia, A., Hudson, N., Foster, I., & Stevens, R. (2024). Causal Discovery over High-Dimensional Structured Hypothesis Spaces with Causal Graph Partitioning. arXiv preprint arXiv:2406.06348.
- Nodoushan, A. N. (2023). Interpretability of Deep Learning Models for Time-Series Clinical Data. (Doctoral dissertation, The University of Arizona).
- Saarela, M., & Podgorelec, V. (2024). Recent Applications of Explainable AI (XAI): A Systematic Literature Review. Applied Sciences, 14(19), 8884.
- Nayebi, A., Tipirneni, S., Reddy, C. K., Foreman, B., & Subbian, V. (2023). WindowSHAP: An efficient framework for explaining time-series classifiers based on Shapley values. Journal of Biomedical Informatics, 144, 104438.
| MODULE | DESCRIPTION |
|---|---|
sliding_window |
TemporalScope/src/temporalscope/partition/single_target/sliding_window.py |