Causal feature selection framework for stable soft sensor modeling based on time-delayed cross mapping
- URL: http://arxiv.org/abs/2601.14099v1
- Date: Tue, 20 Jan 2026 15:58:51 GMT
- Title: Causal feature selection framework for stable soft sensor modeling based on time-delayed cross mapping
- Authors: Shi-Shun Chen, Xiao-Yang Li, Enrico Zio,
- Abstract summary: Causal feature selection can enhance the performance of soft sensor models in industrial applications.<n>This paper proposes a causal feature selection framework based on time-delayed cross mapping.
- Score: 8.141412943138107
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Soft sensor modeling plays a crucial role in process monitoring. Causal feature selection can enhance the performance of soft sensor models in industrial applications. However, existing methods ignore two critical characteristics of industrial processes. Firstly, causal relationships between variables always involve time delays, whereas most causal feature selection methods investigate causal relationships in the same time dimension. Secondly, variables in industrial processes are often interdependent, which contradicts the decorrelation assumption of traditional causal inference methods. Consequently, soft sensor models based on existing causal feature selection approaches often lack sufficient accuracy and stability. To overcome these challenges, this paper proposes a causal feature selection framework based on time-delayed cross mapping. Time-delayed cross mapping employs state space reconstruction to effectively handle interdependent variables in causality analysis, and considers varying causal strength across time delay. Time-delayed convergent cross mapping (TDCCM) is introduced for total causal inference, and time-delayed partial cross mapping (TDPCM) is developed for direct causal inference. Then, in order to achieve automatic feature selection, an objective feature selection strategy is presented. The causal threshold is automatically determined based on the model performance on the validation set, and the causal features are then selected. Two real-world case studies show that TDCCM achieves the highest average performance, while TDPCM improves soft sensor stability and performance in the worst scenario. The code is publicly available at https://github.com/dirge1/TDPCM.
Related papers
- Can Causality Cure Confusion Caused By Correlation (in Software Analytics)? [4.082216579462797]
Symbolic models, particularly decision trees, are widely used in software engineering for explainable analytics.<n>Recent studies in software engineering show that both correlational models and causal discovery algorithms suffer from pronounced instability.<n>This study investigates causality-aware split criteria into symbolic models to improve their stability and robustness.
arXiv Detail & Related papers (2026-02-17T23:35:50Z) - Towards Remote Sensing Change Detection with Neural Memory [61.39582645714727]
ChangeTitans is a Titans-based framework for remote sensing change detection.<n>First, we propose VTitans, which integrates neural memory with segmented local attention.<n>Second, we present a hierarchical VTitans-Adapter to refine multi-scale features across different network layers.<n>Third, we introduce TS-CBAM, a two-stream fusion module, to suppress pseudo-changes and enhance detection accuracy.
arXiv Detail & Related papers (2026-02-11T03:50:51Z) - Multiresolution Analysis and Statistical Thresholding on Dynamic Networks [49.09073800467438]
ANIE (Adaptive Network Intensity Estimation) is a multi-resolution framework designed to automatically identify the time scales at which network structure evolves.<n>We show that ANIE adapts to the appropriate time resolution and is able to capture sharp structural changes while remaining robust to noise.
arXiv Detail & Related papers (2025-06-01T22:55:55Z) - Spatial-Temporal Bearing Fault Detection Using Graph Attention Networks and LSTM [0.7864304771129751]
This paper introduces a novel method that combines Graph Attention Network (GAT) and Long Short-Term Memory (LSTM) networks.
This approach captures both spatial and temporal dependencies within sensor data, improving the accuracy of bearing fault detection.
arXiv Detail & Related papers (2024-10-15T12:55:57Z) - Temporal Feature Matters: A Framework for Diffusion Model Quantization [105.3033493564844]
Diffusion models rely on the time-step for the multi-round denoising.<n>We introduce a novel quantization framework that includes three strategies.<n>This framework preserves most of the temporal information and ensures high-quality end-to-end generation.
arXiv Detail & Related papers (2024-07-28T17:46:15Z) - Causality-driven Sequence Segmentation for Enhancing Multiphase Industrial Process Data Analysis and Soft Sensing [4.420321822469078]
This article introduces a causality-driven sequence segmentation model.
It segments the sequence into different phases based on the sudden shifts in causal mechanisms that occur during phase transitions.
A soft sensing model called temporal-causal graph convolutional network (TC-GCN) is trained for each phase.
arXiv Detail & Related papers (2024-06-30T10:40:54Z) - On the Identification of Temporally Causal Representation with Instantaneous Dependence [50.14432597910128]
Temporally causal representation learning aims to identify the latent causal process from time series observations.
Most methods require the assumption that the latent causal processes do not have instantaneous relations.
We propose an textbfIDentification framework for instantanetextbfOus textbfLatent dynamics.
arXiv Detail & Related papers (2024-05-24T08:08:05Z) - Causal Feature Selection via Transfer Entropy [59.999594949050596]
Causal discovery aims to identify causal relationships between features with observational data.
We introduce a new causal feature selection approach that relies on the forward and backward feature selection procedures.
We provide theoretical guarantees on the regression and classification errors for both the exact and the finite-sample cases.
arXiv Detail & Related papers (2023-10-17T08:04:45Z) - Hawkes Processes with Delayed Granger Causality [9.664517084506718]
We explicitly model the delayed Granger causal effects based on multivariate Hawkes processes.
We infer the posterior distribution of the time lags and understand how this distribution varies across different scenarios.
We empirically evaluate our model's event prediction and time-lag inference accuracy on synthetic and real data.
arXiv Detail & Related papers (2023-08-11T12:43:43Z) - Selecting Robust Features for Machine Learning Applications using
Multidata Causal Discovery [7.8814500102882805]
We introduce a Multidata causal feature selection approach that simultaneously processes an ensemble of time series datasets.
This approach uses the causal discovery algorithms PC1 or PCMCI that are implemented in the Tigramite Python package.
We apply our framework to the statistical intensity prediction of Western Pacific Tropical Cyclones.
arXiv Detail & Related papers (2023-04-11T15:43:34Z) - Deconfounded Video Moment Retrieval with Causal Intervention [80.90604360072831]
We tackle the task of video moment retrieval (VMR), which aims to localize a specific moment in a video according to a textual query.
Existing methods primarily model the matching relationship between query and moment by complex cross-modal interactions.
We propose a causality-inspired VMR framework that builds structural causal model to capture the true effect of query and video content on the prediction.
arXiv Detail & Related papers (2021-06-03T01:33:26Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.