Nonlinear Interdependence Research Articles

Integrated Systematic Framework for Forecasting China’s Consumer Confidence: A Machine Learning Approach

This study aims to introduce a novel approach for predicting China’s consumer confidence index (CCI), a key economic indicator that reflects consumers’ confidence in current and future economic conditions. While traditional statistical models and economic indicators are the primary tools for forecasting CCI, their reliance on linear assumptions limits their ability to capture the complex, dynamic relationships inherent in economic systems. In response, this study proposes a two-step method that integrates social network analysis (SNA) and machine learning (ML) to enhance prediction accuracy by accounting for the nonlinear interactions and systemic interdependencies that drive consumer confidence. The use of SNA enables the identification of critical variables and their interconnected roles in shaping consumer sentiment, while ML models, specifically the gradient boosting decision tree (GBDT), leverage these relationships to provide more precise predictions. Utilizing monthly data from 1999 to 2023, the combined SNA and GBDT approach significantly improves the accuracy of CCI forecasts, particularly during periods of high volatility. The results of this study hold substantial value for policymakers, market analysts, and economists, as they offer a systems-oriented framework for economic forecasting. By demonstrating the effectiveness of combining SNA with ML technologies, this research not only advances the methodological toolkit for economic forecasting, but also provides a new lens through which the complex, adaptive nature of economic systems can be better understood and managed. This integrated approach paves the way for future developments in forecasting models that more accurately reflect the evolving dynamics of consumer confidence in a rapidly changing economic environment.

Building cybersecurity resilience: integrating defense and recovery investment strategies in an expected resilience framework

PurposeTo remain competitive in an unpredictable environment where the complexity and frequency of cybercrime are rapidly increasing, a cyber resiliency strategy is vital for business continuity. However, one of the barriers to improving cyber resilience is that security defense and accident recovery do not combine efficaciously, as embodied by emphasizing cyber security defense strategies, leaving firms ill-prepared to respond to attacks. The present study thus develops an expected resilience framework to assess cyber resilience, analyze cyber security defense and recovery investment strategies and balance security investment allocation strategies.Design/methodology/approachBased on the expected utility theory, this paper presents an expected resilience framework, including an expected investment resilience model and an expected profit resilience model that directly addresses the optimal joint investment decisions between defense and recovery. The effects of linear and nonlinear recovery functions, risk interdependence and cyber insurance on defense and recovery investment are also analyzed.FindingsAccording to the findings, increasing the defense investment coefficient reduces defense and recovery investment while increasing the expected resilience. The nonlinear recovery function requires a smaller defense investment and overall security investment than the linear one, reflecting the former’s advantages in lowering cybersecurity costs. Moreover, risk interdependence has positive externalities for boosting defense and recovery investment, meaning that the expected profit resilience model can reduce free-riding behavior in security investments. Insurance creates moral hazard for firms by lowering defensive investment, yet after purchasing insurance, expanded coverage and cost-effectiveness incentivize firms to increase defense and recovery spending, respectively.Originality/valueThe paper is innovative in its methodology as it offers an expected cyber resilience framework for integrating defense and recovery investment and their effects on security investment allocation, which is crucial for building cybersecurity resilience but receives little attention in cybersecurity economics. It also provides theoretical advances for cyber resilience assessment and optimum investment allocation in other fields, such as cyber-physical systems, power and water infrastructure – moving from a resilience triangle metric to an expected utility theory-based method.

Functional Hypergraphs of Stock Markets.

In stock markets, nonlinear interdependencies between various companies result in nontrivial time-varying patterns in stock prices. A network representation of these interdependencies has been successful in identifying and understanding hidden signals before major events like stock market crashes. However, these studies have revolved around the assumption that correlations are mediated in a pairwise manner, whereas, in a system as intricate as this, the interactions need not be limited to pairwise only. Here, we introduce a general methodology using information-theoretic tools to construct a higher-order representation of the stock market data, which we call functional hypergraphs. This framework enables us to examine stock market events by analyzing the following functional hypergraph quantities: Forman-Ricci curvature, von Neumann entropy, and eigenvector centrality. We compare the corresponding quantities of networks and hypergraphs to analyze the evolution of both structures and observe features like robustness towards events like crashes during the course of a time period.

Quantifying Predictive Uncertainty and Feature Selection in River Bed Load Estimation: A Multi-Model Machine Learning Approach with Particle Swarm Optimization

This study presents a comprehensive multi-model machine learning (ML) approach to predict river bed load, addressing the challenge of quantifying predictive uncertainty in fluvial geomorphology. Six ML models—random forest (RF), categorical boosting (CAT), extra tree regression (ETR), gradient boosting machine (GBM), Bayesian regression model (BRM), and K-nearest neighbors (KNNs)—were thoroughly evaluated across several performance metrics like root mean square error (RMSE), and correlation coefficient (R). To enhance model training and optimize performance, particle swarm optimization (PSO) was employed for hyperparameter tuning across all the models, leveraging its capability to efficiently explore complex hyperparameter spaces. Our findings indicated that RF, GBM, CAT, and ETR demonstrate superior predictive performance (R score > 0.936), benefiting significantly from PSO. In contrast, BRM displayed lower performance (0.838), indicating challenges with Bayesian approaches. The feature importance analysis, including permutation feature and SHAP values, highlighted the non-linear interdependencies between the variables, with river discharge (Q), bed slope (S), and flow width (W) being the most influential. This study also examined the specific impact of individual variables on model performance by adding and excluding individual variables, which is particularly meaningful when choosing input variables for the model, especially in limited data conditions. Uncertainty quantification through Monte Carlo simulations highlighted the enhanced predictability and reliability of models with larger datasets. The correlation between increased training data and improved model precision was evident in the consistent rise in mean R scores and reduction in standard deviations as the sample size increased. This research underscored the potential of advanced ensemble methods and PSO to mitigate the limitations of single-predictor models and exploit collective model strengths, thereby improving the reliability of predictions in river bed load estimation. The insights from this study provide a valuable framework for future research directions focused on optimizing ensemble configurations for hydro-dynamic modeling.

Using qualitative system dynamics modeling to understand overdose bystander behavior in the context of Connecticut’s Good Samaritan Laws and identify effective policy options

BackgroundGood Samaritan Laws are a harm reduction policy intended to facilitate a reduction in fatal opioid overdoses by enabling bystanders, first responders, and health care providers to assist individuals experiencing an overdose without facing civil or criminal liability. However, Good Samaritan Laws may not be reaching their full impact in many communities due to a lack of knowledge of protections under these laws, distrust in law enforcement, and fear of legal consequences among potential bystanders. The purpose of this study was to develop a systems-level understanding of the factors influencing bystander responses to opioid overdose in the context of Connecticut’s Good Samaritan Laws and identify high-leverage policies for improving opioid-related outcomes and implementation of these laws in Connecticut (CT).MethodsWe conducted six group model building (GMB) workshops that engaged a diverse set of participants with medical and community expertise and lived bystander experience. Through an iterative, stakeholder-engaged process, we developed, refined, and validated a qualitative system dynamics (SD) model in the form of a causal loop diagram (CLD).ResultsOur resulting qualitative SD model captures our GMB participants’ collective understanding of the dynamics driving bystander behavior and other factors influencing the effectiveness of Good Samaritan Laws in the state of CT. In this model, we identified seven balancing (B) and eight reinforcing (R) feedback loops within four narrative domains: Narrative 1 - Overdose, Calling 911, and First Responder Burnout; Narrative 2 - Naloxone Use, Acceptability, and Linking Patients to Services; Narrative 3 - Drug Arrests, Belief in Good Samaritan Laws, and Community Trust in Police; and Narrative 4 - Bystander Naloxone Use, Community Participation in Harm Reduction, and Cultural Change Towards Carrying Naloxone.ConclusionsOur qualitative SD model brings a nuanced systems perspective to the literature on bystander behavior in the context of Good Samaritan Laws. Our model, grounded in local knowledge and experience, shows how the hypothesized non-linear interdependencies of the social, structural, and policy determinants of bystander behavior collectively form endogenous feedback loops that can be leveraged to design policies to advance and sustain systems change.

Beyond built environment: Unveiling the interplay of streetscape perceptions and cycling behavior

As an important means of shared micro-mobility, shared bicycles have become a crucial component of urban transportation in China. The impact of the built environment on bicycling has been widely acknowledged. However, can streetscape perceptions influence bicycle-sharing volume (BSV) and supplement the built environment? We first obtained millions of pieces of shared-cycling data from the Shenzhen Open Data Platform and carried out geographical quantification of BSV. As for streetscape, we improved the classification of subjective streetscape perception based on street view images using the k-means clustering algorithm and conducted predictions using XGBoost. Through the application of different regression models, we unveiled the nonlinear spatial interdependencies between BSV and streetscape perceptions as a complement to the built environment. Our findings indicate that greenery, vivid street-front facades, and diverse street facilities can promote BSV. Targeted strategies are proposed for different districts. For instance, urban planners can provide incentives for high-income groups in central urban areas to adopt active travel, and increase the supply of shared bicycles in suburban areas with high building density, particularly in industrial urban villages. As a supplement to the long-term planning recommendations derived from the macro-built environment analysis, an in-depth spatial perception quantitative assessment proffers a human-centric, flexible blueprint for urban street design.

The role of flexural particles in the shear flow of pine residue biomass: An experiment-informed DEM simulation study

An experiment-informed discrete particle simulation study was performed to investigate the role of flexural particles in the shear flow of milled pine residues. Physical pine anatomical fraction samples were tested in an FT4 powder rheometer. A chip model for bark/stem chips and a fiber model for needles were proposed for DEM. The FT4 simulations show mixed fibers/chips require higher force and torque to drive shear flow than pure chips or fibers. The mass fraction and flexibility of fibers reveal nonlinear inter-dependency on impacting the shear flow of mixed fibers/chips. Polydisperse fibers well match the needle-rich samples on the flow attributes, while polydisperse mixed fibers/chips agree reasonably with the bark/stem-rich samples. The DEM study indicates that the flowability of pine residue biomass is determined collectively by various material attributes of the anatomical fractions, which must be considered when developing preprocessing methods for conversion-ready biomass with adequate material handling performance.

Mutual information between Polish subindexes – the use of copula entropy around the time of the COVID-19 pandemic

In this paper, the copula theory is used to describe the dependence structure between variables, while the information theory provides the tools necessary to measure the uncertainty associated with these variables. What both theories have in common is copula entropy, which is strictly related to mutual information. The findings of this study, focusing on the dependence of the (sub)indexes of the Polish stock market during the pandemic period, may prove useful not only to investors from Poland, but also from other countries, especially Central European, in making investment decisions. The results of calculating the interdependencies between WIG, sectoral indexes and among sectoral indexes of the Polish economy using copula entropy and Pearson’s correlation are quite different. The source of the basic difference between copula entropy and Pearson’s correlation is that the former enables the measurement of nonlinear interdependencies, while the latter not. The interrelations on the stock markets are nonlinear and returns are not normally distributed in general. The use of copulas is also superior in terms of ranking correlation, as it is more general and allows the examination of the structure of dependencies between extreme values.

Learning Unified Anchor Graph for Joint Clustering of Hyperspectral and LiDAR Data.

The joint clustering of multimodal remote sensing (RS) data poses a critical and challenging task in Earth observation. Although recent advances in multiview subspace clustering have shown remarkable success, existing methods become computationally prohibitive when dealing with large-scale RS datasets. Moreover, they neglect intrinsic nonlinear and spatial interdependencies among heterogeneous RS data and lack generalization ability for out-of-sample data, thereby restricting their applicability. This article introduces a novel unified framework called anchor-based multiview kernel subspace clustering with spatial regularization (AMKSC). It learns a scalable anchor graph in the kernel space, leveraging contributions from each modality instead of seeking a consensus full graph in the feature space. To ensure spatial consistency, we incorporate a spatial smoothing operation into the formulation. The method is efficiently solved using an alternating optimization strategy, and we provide theoretical evidence of its scalability with linear computational complexity. Furthermore, an out-of-sample extension of AMKSC based on multiview collaborative representation-based classification is introduced, enabling the handling of larger datasets and unseen instances. Extensive experiments on three real heterogeneous RS datasets confirm the superiority of our proposed approach over state-of-the-art methods in terms of clustering performance and time efficiency. The source code is available at https://github.com/AngryCai/AMKSC.

Short-term wind power forecasting based on multivariate/multi-step LSTM with temporal feature attention mechanism

Precision enhancement for short-term wind power forecasting can alleviate negative impact of the forecasting results on wind power generation. Due to complexities and nonlinearities among factors and facets in wind power, it is essential to achieve reliable and stable power generation via the long short-term memory (LSTM) forecasting. To this purpose, multi-task temporal feature attention (MTTFA) based LSTM, namely MTTFA-LSTM, is proposed for multivariate/multi-step wind power forecasting with historical power and meteorological data, in which task-sharing and task-specifying layers are designed for task co-features extracting and task specifics discriminating, respectively. More specifically, in the task-sharing layer, multi-dimensional inputs are fed into LSTM to extract long-term trends, while in the task-specifying layer, one-dimensional convolution operations extract temporal features hidden in each and all time steps. Furthermore, an attention mechanism is adopted to adaptively tune weights for temporal features. Finally, the proposed model is leveraged to cope with different short-term wind power forecasting (SWPF) problems based on the national renewable energy laboratory’s (NREL) wind power data. Simulation results show that the proposed MTTFA-LSTM achieves persistent excellent forecasting accuracy, comparing its backbone STL model, TFA-LSTM as well as the benchmark MTL models in the same setting, which indicate that the complex and non-linear interdependencies among multi-dimensional data can be well depicted by the proposed model.

Learning and integration of adaptive hybrid graph structures for multivariate time series forecasting

Recent status-of-the-art methods for multivariate time series forecasting can be categorized into graph-based approach and global-local approach. The former approach uses graphs to represent the dependencies among variables and apply graph neural networks to the forecasting problem. The latter approach decomposes the matrix of multivariate time series into global components and local components to capture the shared information across variables. However, both approaches cannot capture the propagation delay of the dependencies among individual variables of a multivariate time series, for example, the congestion at intersection A has delayed effects on the neighboring intersection B. In addition, graph-based forecasting methods cannot capture the shared global tendency across the variables of a multivariate time series; and global-local forecasting methods cannot reflect the nonlinear inter-dependencies among variables of a multivariate time series. In this paper, we propose to combine the advantages of both approaches by integrating Adaptive Global-Local Graph Structure Learning with Gated Recurrent Units (AGLG-GRU). We learn a global graph to represent the shared information across variables. And we learn dynamic local graphs to capture the local randomness and nonlinear dependencies among variables. We apply diffusion convolution and graph convolution operations to global and dynamic local graphs to integrate the information of graphs and update gated recurrent unit for multivariate time series forecasting. The experimental results on seven representative real-world datasets demonstrate that our approach outperforms various existing methods.

Frequency domain causality and quantile connectedness between investor sentiment and cryptocurrency returns

This study investigates the frequency-domain causality and quantile connectedness between online investors’ fear sentiment and cryptocurrency returns. We propose cross-quantile coherency and networks to examine the frequency-domain nonlinear interdependence. First, we find that investor fear sentiment and cryptocurrency returns exhibit bidirectional causality. Second, fear exhibits an asymmetric connectedness with cryptocurrency returns across quantiles and frequencies. Third, short-term cross-quantile connectedness is found to be more significant than long-term connectedness. These findings can help investors and policymakers make decisions regarding diversified hedging and controlling for potential risks.

Causal interaction in high frequency turbulence at the biosphere-atmosphere interface: Structure-function coupling.

At the biosphere-atmosphere interface, nonlinear interdependencies among components of an ecohydrological complex system can be inferred using multivariate high frequency time series observations. Information flow among these interacting variables allows us to represent the causal dependencies in the form of a directed acyclic graph (DAG). We use high frequency multivariate data at 10 Hz from an eddy covariance instrument located at 25 m above agricultural land in the Midwestern US to quantify the evolutionary dynamics of this complex system using a sequence of DAGs by examining the structural dependency of information flow and the associated functional response. We investigate whether functional differences correspond to structural differences or if there are no functional variations despite the structural differences. We base our analysis on the hypothesis that causal dependencies are instigated through information flow, and the resulting interactions sustain the dynamics and its functionality. To test our hypothesis, we build upon causal structure analysis in the companion paper to characterize the information flow in similarly clustered DAGs from 3-min non-overlapping contiguous windows in the observational data. We characterize functionality as the nature of interactions as discerned through redundant, unique, and synergistic components of information flow. Through this analysis, we find that in turbulence at the biosphere-atmosphere interface, the variables that control the dynamic character of the atmosphere as well as the thermodynamics are driven by non-local conditions, while the scalar transport associated with CO2 and H2O is mainly driven by short-term local conditions.

Nonlinear interdependence of the results of measuring anti-SARS-CoV-2 IgG levels using Abbott and Euroimmun test systems.

There are significant number of tests used to determine the level of antibodies to SARS-CoV-2 which differ both in the methods underlying testing and in the antigenic targets used and classes of measured immunoglobulins. Comparison of the results obtained using various tests reveals their significant discrepancy when converted to the WHO recommended standard unit for measuring the level of specific immunoglobulins BAU/mL. The aim of this study is a comparison of anty-SARS-CoV-2 IgG levels, measured using test systems based on different methodological platforms - EuroImmun assay and Abbott assay. Abbott uses the immunochemiluminescence method CLIA, EuroImmun uses the enzyme immunoassay method ELISA. The dependences of the measurement error on the level of antibodies for the two test systems were approximated by power functions using the least squares method. The nonlinear relation of antibody levels values measured by Abbott assay and Euroimmun assay was approximated by an asymptotic function. The study involved 112 people. Our results confirm the fallacy of using a single conversion coefficient in BAU/mL for anti-SARS-CoV-2 IgG levels measured by Abbott and EuroImmun. To describe the interdependence of anti-SARS-CoV-2 IgG Abbott and EuroImmun levels, we offer the function y=18/π arctan(0.0009x) and a calculator that allows to easily recalculate the results obtained using these tests. The non-linear nature of the interdependence of the measured anti-SARS-CoV-2 antibodies levels on the levels magnitude is one of the main reasons for the discrepancy between the tests results when converted to BAU/mL using a single conversion coefficient.

Dynamic Resilience Quantification of Hydropower Infrastructure in Multihazard Environments

Ensuring the continued functionality of hydropower infrastructure is of the greatest importance, considering the devastating socioeconomic and environmental impacts of dam operation failures. Among the different approaches currently adopted in hydropower dam operational safety, those that are resilience-based are at the leading edge because they focus on assessing the dynamic system performance pre-, during-, and post-hazard exposures. However, the main challenge for such assessment pertains to the complexity associated with the dynamic operation simulation of hydropower dam systems that consist of several components with nonlinear interdependencies. Moreover, the infrastructure’s exposure to a multihazard environment, which may impact one or more hydropower critical system components, poses further challenges to understanding possible subsequent dam operation failure scenarios. This study develops a resilience-centric system dynamics simulation model that provides a holistic representation of hydropower dam system components to estimate the system’s dynamic resilience in multihazard environments. The study also discusses a combinatorial procedure to generate multihazard scenarios, where a primary hazard can trigger one or more subsequent hazards. Finally, an actual hydropower dam is employed to demonstrate the developed model utility in assessing the resilience of complex infrastructure under a wide range of multihazard scenarios. The proposed model provides valuable decision support tools for infrastructure systemic risk mitigation in multihazard environments—facilitating the development of effective resilience planning strategies.

Multi-scale spectral-spatial dual-transformer network for hyperspectral image classification

ABSTRACT Deep learning methods have shown great advantages in hyperspectral image (HSI) classification tasks. In particular, convolutional neural network (CNN)-based methods for HSI classification have made great progress. However, it is difficult for CNN to process long-range spatial and spectral information. Recently, a novel deep learning model called transformer has demonstrated its potential to replace CNN in various classification tasks with its amazing performance. In this letter, a multi-scale spectral-spatial dual-transformer network (MS3DT) is proposed to deeply consider the spectral-spatial features via transformer. Specifically, MS3DT consists of a feature pyramid network (FPN), a spectral transformer subnetwork (SPECT) and a spatial transformer subnetwork (SPAT). To utilize the complementary multi-scale characteristics, we introduce FPN to capture shallow-to-deep and spectral-spatial features. To improve the representational capacity in spatial and spectral domains, SPECT is exploited to extract long-range spectral correlation over local spectral features, and SPAT is designed to explore the contextual information for better refinement. Therefore, MS3DT is able to adaptively recalibrate the nonlinear interdependence of shallow-to-deep spectral-spatial features by merging the two subnetworks. Experiments on two widely used HSI datasets show that the MS3DT can outperform the state-of-the-art (SOTA) algorithm. The source codes will be available at https://github.com/RsAI-lab/MSSSDT.

Quantifying bus route service disruptions under interdependent cascading failures of a multimodal public transit system based on an improved coupled map lattice model

The coupled map lattice (CML) model can present rich spatiotemporal dynamic behaviors of complex systems, e.g., it has been observed to have good adaptability in the cascading failure modeling of subway networks (SNs). However, most studies rarely considered an interdependency between an SN and a bus network (BN), which reveals an open problem concerning the CML model's adaptability to a multimodal public transit system (MPTS). This paper develops an improved CML model to simulate interdependent cascading failures of an MPTS and quantify bus route service disruptions for amplifying severe consequences of failures. Remarkably, an interdependency integrating operational and nonlinear geographical interdependencies is proposed to modify the node state evolutionary function and model cascading failures across layers. Moreover, an estimation method for the node initial state is proposed to accommodate the initial value-sensitive dependence due to a logistic chaotic map involved in the evolutionary function. Finally, a case is simulated to verify the model's adaptability and enlighten operation management. Results indicate that (i) the neglect of interdependency causes an overestimation of the severity of cascading failures of an SN while a severe underestimation of those of a BN; (ii) the management and engineering measures to adjust transfers both have a phase transition point on the controllability of cascading failures, and the latter exhibits more a direct and comprehensive effect.

Forecasting Short-Term Passenger Flow of Subway Stations Based on the Temporal Pattern Attention Mechanism and the Long Short-Term Memory Network

Rational use of urban underground space (UUS) and public transportation transfer underground can solve urban traffic problems. Accurate short-term prediction of passenger flow can ensure the efficient, safe, and comfortable operation of subway stations. However, complex and nonlinear interdependencies between time steps and time series complicate such predictions. This study considered temporal patterns across multiple time steps and selected relevant information on short-term passenger flow for prediction. A hybrid model based on the temporal pattern attention (TPA) mechanism and the long short-term memory (LSTM) network was developed (i.e., TPA-LSTM) for predicting the future number of passengers in subway stations. The TPA mechanism focuses on the hidden layer output values of different time steps in history and of the current time as well as correlates these output values to improve the accuracy of the model. The card swiping data from the Hangzhou Metro automatic fare collection system in China were used for verification and analysis. This model was compared with a convolutional neural network (CNN), LSTM, and CNN-LSTM. The results showed that the TPA-LSTM outperformed the other models with good applicability and accuracy. This study provides a theoretical basis for the pre-allocation of subway resources to avoid subway station crowding and stampede accidents.

Adaptive Mask Sampling and Manifold to Euclidean Subspace Learning With Distance Covariance Representation for Hyperspectral Image Classification

For the abundant spectral and spatial information recorded in hyperspectral images (HSIs), fully exploring spectral-spatial relationships has attracted widespread attention in hyperspectral image classification (HSIC) community. However, there are still some intractable obstructs. For one thing, in the patch based processing pattern, some spatial neighbor pixels are often inconsistent with the central pixel in land-cover class. For another thing, linear and nonlinear correlations between different spectral bands are vital yet tough for representing and excavating. To overcome these mentioned issues, an adaptive mask sampling and manifold to Euclidean subspace learning (AMS-M2ESL) framework is proposed for HSIC. Specifically, an adaptive mask based intra-patch sampling (AMIPS) module is firstly formulated for intra-patch sampling in an adaptive mask manner based on central spectral vector oriented spatial relationships. Then, based on distance covariance descriptor, a dual channel distance covariance representation (DC-DCR) module is proposed for modeling unified spectral-spatial feature representations and exploring spectral-spatial relationships, especially linear and nonlinear interdependence in spectral domain. Furthermore, considering that distance covariance matrix lies on the symmetric positive definite (SPD) manifold, we implement a manifold to Euclidean subspace learning (M2ESL) module respecting Riemannian geometry of SPD manifold for high-level spectral-spatial feature learning. Additionally, we introduce an approximate matrix square-root (ASQRT) layer for efficient Euclidean subspace projection. Extensive experimental results on three popular HSI data sets with limited training samples demonstrate the superior performance of the proposed method compared with other state-of-the-art methods. The source code is available at https://github.com/lms-07/AMS-M2ESL.

Cause-Effect Relationships in Battery Cell Production - Data based validation of expert knowledge in electrode production

The advancing climate change requires a shift from conventional energy sources to renewable ones, which is accompanied by the indispensability of an efficient and sustainable storage of electrical energy. The production of lithium-ion battery cells comprises an enormous complexity, as it consists of several interlinked process steps, each with different quality requirements for the (intermediate) products and the processes itself. This complexity means that a supposedly insignificant process parameter can lead to errors and deviations in the manufactured products via complex cause-effect relationships, that are not fully known. The missing knowledge of the processes and inter-process interdependencies lead to difficulties in the setup of a suitable quality assurance in battery cell production to reduce scrap rates. Beside these potentials in planning and optimizing the process chain of the battery cell factory of tomorrow are impeded.We developed a method to build up a relevant knowledge base on cause-effect relationships for battery cell production. Expert knowledge based on experience and intuition has been collected and quantitative cause-effect relationships are established. This knowledge is used to set up the initial data acquisition, ensuring relevant data being available for analysis. Afterwards, the data analysis, that validates and extends this expert knowledge, is defined. This approach is applied to real-world use-cases including machinery and laboratory equipment in battery cell production. The electrode production is selected as the basis for validation. The found inter-relationships either validate the qualitative expert knowledge, contradict this prior knowledge or else add new insights about the battery cell electrode production process that have not been considered in the qualitative expert knowledge. Methods used include correlations and a predictive power score for non-linear interdependencies. The cause-effect relationships are constantly updated by automatically analyzing the production data of the digital factory. These identified and validated cause-effect relationships allow the planning of efficient and sustainable battery cell production and data-based process chain optimization.