Modeling cancer progression: an integrated workflow extending data-driven kinetic models to bio-mechanical PDE models.

Computational modeling of cancer can help unveil dynamics and interactions that are hard to replicate experimentally. Thanks to the advancement in cancer databases and data analysis technologies, these models have become more robust than ever. There are many mathematical models which investigate cancer through different approaches, from sub-cellular to tissue scale, and from treatment to diagnostic points of view. In this study, we lay out a step-by-step methodology for a data-driven mechanistic model of the tumor microenvironment. We discuss data acquisition strategies, data preparation, parameter estimation, and sensitivity analysis techniques. Furthermore, we propose a possible approach to extend mechanistic ordinary differential equation models to PDE models coupled with mechanical growth. The workflow discussed in this article can help understand the complex temporal and spatial interactions between cells and cytokines in the tumor microenvironment and their effect on tumor growth.

Physics Constrained High-Precision Data-Driven Modeling for Multi-Path Ultrasonic Flow Meter in Natural Gas Measurement.

Ultrasonic flow meters are crucial measuring instruments in natural gas transportation pipeline scenarios. The collected flow velocity data, along with the operational conditions data, are vital for the analysis of the metering performance of ultrasonic flow meters and analysis of the flow process. In practical applications, high requirements are placed on the modeling accuracy of ultrasonic flow meters. In response, this paper proposes an ultrasonic flow meter modeling method based on a combination of data learning and industrial physics knowledge. This paper builds ultrasonic flow meter flow velocity prediction models under different working conditions, combining pipeline flow field velocity distribution knowledge for data preprocessing and loss function design. By making full use of the characteristics of the physics and data learning, the prediction results are close to the real acoustic path flow velocity distribution; thus, the model has high accuracy and interpretability. Experiments are conducted to prove that the prediction error of the proposed method can be controlled within 1%, which can meet the needs of ultrasonic flow meter modeling and subsequent performance analysis in actual production.

Towards good modelling practice for parallel hybrid models for wastewater treatment processes.

This study explores various approaches to formulating a parallel hybrid model (HM) for Water and Resource Recovery Facilities (WRRFs) merging a mechanistic and a data-driven model. In the study, the HM is constructed by training a neural network (NN) on the residual of the mechanistic model for effluent nitrate. In an initial experiment using the Benchmark Simulation Model no. 1, a parallel HM effectively addressed limitations in the mechanistic model's representation of autotrophic bacteria growth and the data-driven model's incapability to extrapolate. Next, different versions of a parallel HM of a large pilot-scale WRRF are constructed, using different calibration/training datasets and different versions of the mechanistic model to investigate the balance between the calibration effort for the mechanistic model and the compensation by the NN component. The HM can improve predictions compared to the mechanistic model. Training the NN on an independent validation dataset produced better results than on the calibration dataset. Interestingly, the best performance is achieved for the HM based on a mechanistic model using default (uncalibrated) parameters. Both long short-term memory (LSTM) and convolutional neural network (CNN) are tested as data-driven components, with a CNN HM (root-mean-squared error (RMSE) = 1.58 mg NO3-N/L) outperforming an LSTM HM (RMSE = 4.17 mg NO3-N/L).

Data-driven prediction models for forecasting multistep ahead profiles of mammalian cell culture toward bioprocess digital twins.

Recently, the advancement in process analytical technology and artificial intelligence (AI) has enabled the generation of enormous culture data sets from biomanufacturing processes that produce various recombinant therapeutic proteins (RTPs), such as monoclonal antibodies (mAbs). Thus, now it is very important to exploit them for the enhanced reliability, efficiency, and consistency of the RTP-producing culture processes and for the reduced incipient or abrupt faults. It is achievable by AI-based data-driven models (DDMs), which allow us to correlate biological and process conditions and cell culture states. In this work, we provide practical guidelines for choosing the best combination of model elements to design and implement successful DDMs for given hypothetical in-line data sets during mAb-producing Chinese hamster ovary cell culture, as such enabling us to forecast dynamic behaviors of culture performance such as viable cell density, mAb titer as well as glucose, lactate and ammonia concentrations. To do so, we created DDMs that balance computational load with model accuracy and reliability by identifying the best combination of multistep ahead forecasting strategies, input features, and AI algorithms, which is potentially applicable to implementation of interactive DDM within bioprocess digital twins. We believe this systematic study can help bioprocess engineers start developing predictive DDMs with their own data sets and learn how their cell cultures behave in near future, thereby rendering proactive decision possible.

Data-driven models applying in household hazardous waste: Amount prediction and classification in Shanghai.

Precisely predicting the amount of household hazardous waste (HHW) and classifying it intelligently is crucial for effective city management. Although data-driven models have the potential to address these problems, there have been few studies utilizing this approach for HHW prediction and classification due to the scarcity of available data. To address this, the current study employed the prophet model to forecast HHW quantities based on the Integration of Two Networks systems in Shanghai. HHW classification was performed using HVGGNet structures, which were based on VGG and transfer learning. To expedite the process of finding the optimal global learning rate, the method of cyclical learning rate was adopted, thus avoiding the need for repeated testing. Results showed that the average rate of HHW generation was 0.1 g/person/day, with the most significant waste categories being fluorescent lamps (30.6 %), paint barrels (26.1 %), medicine (26.2 %), battery (15.8 %), thermometer (0.03 %), and others (1.22 %). Recovering rare earth element (18.85 kg), Cd (3064.10 kg), Hg (15643.43 kg), Zn (14239.07 kg), Ag (11805.81 kg), Ni (4956.64 kg) and Li (1081.45 kg) from HHW can help avoid groundwater pollution, soil contamination and air pollution. HVGGNet-11 demonstrated 90.5 % precision and was deemed most suitable for HHW sorting. Furthermore, the prophet model predicted that HHW in Shanghai would increase from 794.43 t in 2020 to 2049.67 t in 2025.

Reduced Order Modeling of Nonlinear Vibrating Multiphysics Microstructures with Deep Learning-Based Approaches.

Micro-electro-mechanical-systems are complex structures, often involving nonlinearites of geometric and multiphysics nature, that are used as sensors and actuators in countless applications. Starting from full-order representations, we apply deep learning techniques to generate accurate, efficient, and real-time reduced order models to be used for the simulation and optimization of higher-level complex systems. We extensively test the reliability of the proposed procedures on micromirrors, arches, and gyroscopes, as well as displaying intricate dynamical evolutions such as internal resonances. In particular, we discuss the accuracy of the deep learning technique and its ability to replicate and converge to the invariant manifolds predicted using the recently developed direct parametrization approach that allows the extraction of the nonlinear normal modes of large finite element models. Finally, by addressing an electromechanical gyroscope, we show that the non-intrusive deep learning approach generalizes easily to complex multiphysics problems.

Prognostic and Health Management of Critical Aircraft Systems and Components: An Overview.

Prognostic and health management (PHM) plays a vital role in ensuring the safety and reliability of aircraft systems. The process entails the proactive surveillance and evaluation of the state and functional effectiveness of crucial subsystems. The principal aim of PHM is to predict the remaining useful life (RUL) of subsystems and proactively mitigate future breakdowns in order to minimize consequences. The achievement of this objective is helped by employing predictive modeling techniques and doing real-time data analysis. The incorporation of prognostic methodologies is of utmost importance in the execution of condition-based maintenance (CBM), a strategic approach that emphasizes the prioritization of repairing components that have experienced quantifiable damage. Multiple methodologies are employed to support the advancement of prognostics for aviation systems, encompassing physics-based modeling, data-driven techniques, and hybrid prognosis. These methodologies enable the prediction and mitigation of failures by identifying relevant health indicators. Despite the promising outcomes in the aviation sector pertaining to the implementation of PHM, there exists a deficiency in the research concerning the efficient integration of hybrid PHM applications. The primary aim of this paper is to provide a thorough analysis of the current state of research advancements in prognostics for aircraft systems, with a specific focus on prominent algorithms and their practical applications and challenges. The paper concludes by providing a detailed analysis of prospective directions for future research within the field.

Multiple-layer statistical methodology for developing data-driven models of anaerobic digestion process.

When modelling anaerobic digestion, ineffective data handling and inadequate designation of modelling parameters can undermine the model reliability. In this study, a multilayer statistical technique, which employed a machine learning technique using regression models, was introduced to systematically support the development of anaerobic digestion models. Layer-by-layer statistical techniques including cubic smoothing splines (missing data reconstruction), principal component analysis (identifying correlated parameters), analysis of variance (analysing differences among datasets), and linear regression (developing data-driven models) were used to develop and validate anaerobic digestion models. Experimental data collected from the long-term operation of lab-scale (operated for 350 days), pilot-scale (operated for 150 days), and full-scale reactors (operated for 750 days) were used to demonstrate the modelling process. The multivariate models based on a data-driven modelling technique were developed by subjecting the experimental and monitored data to a modelling process. The developed models could predict the biogas production and effluent chemical oxygen demand during anaerobic digestion. Statistical analyses verified the modelling hypotheses, evaded invalid model development, and ensured data integrity and parameter validity. Multiple linear regression of principal components demonstrated that the performance of biogas production using food waste was influenced by the variances of the nitrogen and organic concentrations, but not by the chemical oxygen demand to total nitrogen (C/N) ratio. In the validation process, the model developed with lab-scale reactor data showed relatively high accuracy with R2, SSE, and RMSE values of 0.86, 34.45, and 0.72.

An integrated supervision framework to safeguard the urban river water quality supported by ICT and models.

Models and information and communication technology (ICT) can assist in the effective supervision of urban receiving water bodies and drainage systems. Single model-based decision tools, e.g., water quality models and the pollution source identification (PSI) method, have been widely reported in this field. However, a systematic pathway for environmental decision support system (EDSS) construction by integrating advanced single techniques has rarely been reported, impeding engineering applications. This paper presents an integrated supervision framework (UrbanWQEWIS) involving monitoring-early warning-source identification-emergency disposal to safeguard the urban water quality, where the data, model, equipment and knowledge are smoothly and logically linked. The generic architecture, all-in-one equipment and three key model components are introduced. A pilot EDSS is developed and deployed in the Maozhou River, China, with the assistance of environmental Internet of Things (IoT) technology. These key model components are successfully validated via in situ monitoring data and dye tracing experiments. In particular, fluorescence fingerprint-based qualitative PSI and Bayesian-based quantitative PSI methods are effectively coupled, which can largely reduce system costs and enhance flexibility. The presented supervision framework delivers a state-of-the-art management tool in the digital water era. The proposed technical pathway of EDSS development provides a valuable reference for other regions.

WEAK SINDy: GALERKIN-BASED DATA-DRIVEN MODEL SELECTION.

We present a novel weak formulation and discretization for discovering governing equations from noisy measurement data. This method of learning differential equations from data fits into a new class of algorithms that replace pointwise derivative approximations with linear transformations and variance reduction techniques. Compared to the standard SINDy algorithm presented in [S. L. Brunton, J. L. Proctor, and J. N. Kutz, Proc. Natl. Acad. Sci. USA, 113 (2016), pp. 3932-3937], our so-called weak SINDy (WSINDy) algorithm allows for reliable model identification from data with large noise (often with ratios greater than 0.1) and reduces the error in the recovered coefficients to enable accurate prediction. Moreover, the coefficient error scales linearly with the noise level, leading to high-accuracy recovery in the low-noise regime. Altogether, WSINDy combines the simplicity and efficiency of the SINDy algorithm with the natural noise reduction of integration, as demonstrated in [H. Schaeffer and S. G. McCalla, Phys. Rev. E, 96 (2017), 023302], to arrive at a robust and accurate method of sparse recovery.

Data Driven Model-Free Adaptive Control Method for Quadrotor Formation Trajectory Tracking Based on RISE and ISMC Algorithm.

In order to solve the problems of complex dynamic modeling and parameters identification of quadrotor formation cooperative trajectory tracking control, this paper proposes a data-driven model-free adaptive control method for quadrotor formation based on robust integral of the signum of the error (RISE) and improved sliding mode control (ISMC). The leader-follower strategy is adopted, and the leader realizes trajectory tracking control. A novel asymptotic tracking data-driven controller of quadrotor is used to control the system using the RISE method. It is divided into two parts: The inner loop is for attitude control and the outer loop for position control. Both use the RISE method in the loop to eliminate interference and this method only uses the input and output data of the unmanned aerial vehicle(UAV) system and does not rely on any dynamics and kinematics model of the UAV. The followers realize formation cooperative control, introducing adaptive update law and saturation function to improve sliding mode control (SMC), and it eliminates the general SMC algorithm controller design dependence on the mathematical model of the UAV and has the chattering problem. Then, the stability of the system is proved by the Lyapunov method, and the effectiveness of the algorithm and the feasibility of the scheme are verified by numerical simulation. The experimental results show that the designed data-driven model-free adaptive control method for the quadrotor formation is effective and can effectively realize the coordinated formation trajectory tracking control of the quadrotor. At the same time, the design of the controller does not depend on the UAV kinematics and dynamics model, and it has high control accuracy, stability, and robustness.

Turbidity prediction of lake-type raw water using random forest model based on meteorological data: A case study of Tai lake, China.

Turbidity is an indication of water quality and enables the growth of pathogenic microorganisms. For drinking water treatment plants (DWTPs), violent fluctuations in turbidity are highly disruptive to operational performance due to the lag in process parameter adjustments. Such risks must be carefully managed to guarantee safe drinking water. Machine learning techniques have been proven to be effective for modeling complex nonlinear environmental systems, and this study adopted such a technique to develop a model for predicting source water turbidity for DWTPs to allow DWTPs to make proactive interventions in advance. A random forest (RF) model used preprocessed (empirical mode decomposition and quartile rejecting) meteorological factors (wind speed, wind direction, air temperature, and rainfall) as the input variables, to establish the turbidity prediction of a lake with significant turbidity in China's South Tai Lake. The modeling process included four main stages: (1) source data analysis, (2) raw data preprocessing, (3) modeling and tuning, and (4) model evaluation. The results of the RF model indicated that the correlation coefficient between the predicted and actual sequences is over 0.7, and more than 55% of the predicted values could control the errors within 20% compared to the actual measured values, suggesting that machine learning techniques are suitable for predicting the turbidity of raw source water. It was found that the RF model can provide a modest performance boost because of its stronger capacity to capture nonlinear interactions in the data. The findings of this study can inform the development of turbidity prediction models using readily available meteorological forecast data. The model can be applied to other DWTPs using similar shallow lakes as water sources.

Utilization of random vector functional link integrated with manta ray foraging optimization for effluent prediction of wastewater treatment plant.

An innovative predictive model was employed to predict the key performance indicators of a full-scale wastewater treatment plant (WWTP) operated with an activated sludge treatment process. The data-driven model was obtained using data gathered from Cairo, Egypt. The proposed model consists of Random Vector Functional Link (RVFL) Networks incorporated with Manta Ray Foraging Optimizer (MRFO). RVFL is used as an advanced Artificial Neural Network (ANN) that avoids the common conventional ANN problems such as overfitting. MRFO is employed to determine the best RVFL parameters to maximize the prediction accuracy of the model. The developed MRFO-RVFL is compared with conventional RVFL to figure out the role of MRFO as an optimization tool to enhance model performance. Both models were trained and tested using experimental data measured during a long period of 222 days. This study aims to provide an accurate prediction of the most widely treated effluent indicators of BOD5 and TSS in the wastewater treatment plants. In this study, ten well-known influent wastewater parameters, BOD5, TSS, and VSS, influent flow rate, pH, ambient temperature, F/M ratio, SRT, WAS, and RAS, the output BOD5 and TSS were modeled and predicted using the integrated MRFO-RVFL algorithms and compared with the standalone RVFL model. The performance of the models was evaluated using different assessment measures such as R2, RMSE, and others. The obtained results of R2 and RMSE for the MRFO-RVFL model were 0.924 and 3.528 for BOD5 and 0.917 and 6.153 for TSS, which were much better than the results of conventional RVFL with 0.840 and 6.207 for BOD5 and 0.717 and 10.05 for TSS. Based on the obtained results, the selective model (MRFO-RVFL) exhibited a higher performance and validity to predict the TSS and optimal BOD5.

A Data-Driven Space-Time-Parameter Reduced-Order Model with Manifold Learning for Coupled Problems: Application to Deformable Capsules Flowing in Microchannels.

An innovative data-driven model-order reduction technique is proposed to model dilute micrometric or nanometric suspensions of microcapsules, i.e., microdrops protected in a thin hyperelastic membrane, which are used in Healthcare as innovative drug vehicles. We consider a microcapsule flowing in a similar-size microfluidic channel and vary systematically the governing parameter, namely the capillary number, ratio of the viscous to elastic forces, and the confinement ratio, ratio of the capsule to tube size. The resulting space-time-parameter problem is solved using two global POD reduced bases, determined in the offline stage for the space and parameter variables, respectively. A suitable low-order spatial reduced basis is then computed in the online stage for any new parameter instance. The time evolution of the capsule dynamics is achieved by identifying the nonlinear low-order manifold of the reduced variables; for that, a point cloud of reduced data is computed and a diffuse approximation method is used. Numerical comparisons between the full-order fluid-structure interaction model and the reduced-order one confirm both accuracy and stability of the reduction technique over the whole admissible parameter domain. We believe that such an approach can be applied to a broad range of coupled problems especially involving quasistatic models of structural mechanics.

Real-time in situ auto-correction of K+ interference for continuous and long-term NH4+ monitoring in wastewater using solid-state ion selective membrane (S-ISM) sensor assembly.

Potassium ions (K+) present in wastewater has caused severe interference for NH4+ monitoring, over-estimation of NH4+ concentration and ultimately leads to extra energy consumption. Past effort for enhancing the selectivity of NH4+ over K+ were oftentimes complex, costly, or compromised the selectivity and accuracy of the NH4+ ion selective membrane (ISM) sensors. This study targeted this imminent challenge by developing an integrated NH4+/K+ auto-correction solid-state ISM (S-ISM) sensor assembly combined with a data-driven model to monitor [NH4+] under different [NH4+] and [K+] concentrations. The results showed that the interference of K+ was substantially alleviated for NH4+ measurement. The accuracy was enhanced by over 70% when examined using real wastewater and energy consumption was expected to reduce by 26% for a wastewater treatment plant, especially for wastewater with high [K+]. Furthermore, the uniquely structured S-ISMs were made by embedding the ionophores in a robust polyvinyl chloride (PVC) matrix containing plasticizers and a layer of carbon nanotubes (CNT) as ion-to-electron transducer, which maintained the selectivity and accuracy of the S-ISM sensor for 4 weeks in wastewater. NH4+/K+ sensor assembly integrated with data-driven correction models poses great potential in high-efficiency and energy-saving wastewater treatment and water reuse processes.

Introducing a hybrid artificial intelligence method for high-throughput modeling and optimizing plant tissue culture processes: the establishment of a new embryogenesis medium for chrysanthemum, as a case study.

Data-driven models in a combination of optimization algorithms could be beneficial methods for predicting and optimizing in vitro culture processes. This study was aimed at modeling and optimizing a new embryogenesis medium for chrysanthemum. Three individual data-driven models, including multi-layer perceptron (MLP), adaptive neuro-fuzzy inference system (ANFIS), and support vector regression (SVR), were developed for callogenesis rate (CR), embryogenesis rate (ER), and somatic embryo number (SEN). Consequently, the best obtained results were used in the fusion process by a bagging method. For medium reformulation, effects of eight ionic macronutrients on CR, ER, and SEN and effects of four vitamins on SEN were evaluated using data fusion (DF)-non-dominated sorting genetic algorithm-II (NSGA-II) and DF-genetic algorithm (GA), respectively. Results showed that DF models with the highest R2 had superb performance in comparison with all other individual models. According to DF-NSGAII, the highest ER and SEN can be obtained from the medium containing 14.27 mM NH4+, 38.92 mM NO3-, 22.79 mM K+, 5.08 mM Cl-, 3.34 mM Ca2+, 1.67 mM Mg2+, 2.17 mM SO42-, and 1.44 mM H2PO4-. Based on the DF-GA model, the maximum SEN can be obtained from a medium containing 0.61 µM thiamine, 5.93 µM nicotinic acid, 0.25 µM biotin, and 0.26 µM riboflavin. The efficiency of the established-optimized medium was experimentally compared to Murashige and Skoog medium (MS) for embryogenesis of five chrysanthemum cultivars, and results indicated the efficiency of optimized medium over MS medium.Key points• MLP, SVR, and ANFIS were fused by a bagging method to develop a data fusion model.• NSGA-II and GA were linked to the data fusion model for establishing and optimizing a new embryogenesis medium.• The new culture medium (HNT) had better efficiency than MS medium.

Application of artificial intelligence models and optimization algorithms in plant cell and tissue culture.

Artificial intelligence (AI) models and optimization algorithms (OA) are broadly employed in different fields of technology and science and have recently been applied to improve different stages of plant tissue culture. The usefulness of the application of AI-OA has been demonstrated in the prediction and optimization of length and number of microshoots or roots, biomass in plant cell cultures or hairy root culture, and optimization of environmental conditions to achieve maximum productivity and efficiency, as well as classification of microshoots and somatic embryos. Despite its potential, the use of AI and OA in this field has been limited due to complex definition terms and computational algorithms. Therefore, a systematic review to unravel modeling and optimizing methods is important for plant researchers and has been acknowledged in this study. First, the main steps for AI-OA development (from data selection to evaluation of prediction and classification models), as well as several AI models such as artificial neural networks (ANNs), neurofuzzy logic, support vector machines (SVMs), decision trees, random forest (FR), and genetic algorithms (GA), have been represented. Then, the application of AI-OA models in different steps of plant tissue culture has been discussed and highlighted. This review also points out limitations in the application of AI-OA in different plant tissue culture processes and provides a new view for future study objectives. KEY POINTS: • Artificial intelligence models and optimization algorithms can be considered a novel and reliable computational method in plant tissue culture. • This review provides the main steps and concepts for model development. • The application of machine learning algorithms in different steps of plant tissue culture has been discussed and highlighted.

A data driven epidemic model to analyse the lockdown effect and predict the course of COVID-19 progress in India.

We propose a data driven epidemic model using the real data on the infection, recovery and death cases for the analysis of COVID-19 progression in India. The model assumes continuation of existing control measures such as lockdown and quarantines, the suspected and confirmed cases and does not consider the scenario of 2nd surge of the epidemic due to any reason. The model is arrived after least square fitting of epidemic behaviour model based on theoretical formulation to the real data of cumulative infection cases reported between 24 March 2020 and 30May 2020. The predictive capability of the model has been validated with real data of infection cases reported during June 1-10, 2020. A detailed analysis of model predictions in terms of future trend of COVID-19 progress individually in 18 states of India and India as a whole has been attempted. Infection rate in India, as a whole, is continuously decreasing with time and has reached 3 times lower than the initial infection rate after 6 weeks of lock down suggesting the effectiveness of the lockdown in containing the epidemic. Results suggest that India, as a whole, could see the peak and end of the epidemic in the month of July 2020 and March 2021 respectively as per the current trend in the data. Active infected cases in India may touch 2 lakhs or little above at the peak time and total infected cases may reach over 19 lakhs as per current trend. State-wise results have been discussed in the manuscript. However, the prediction may deviate particularly for longer dates, as assumptions of model cannot be met always in a real scenario. In view of this, a real time application (COV-IND Predictor) has been developed which automatically syncs the latest data from the national COVID19 dash board on daily basis and updates the model input parameters and predictions instantaneously. This real time application can be accessed from the link: https://docs.google.com/spreadsheets/d/1fCwgnQ-dz4J0YWVDHUcbEW1423wOJjdEXm8TqJDWNAk/edit?usp=sharing and can serve as a practical tool for policy makers to track peak time and maximum active infected cases based on latest trend in data for medical readiness and taking epidemic management decisions.

Data-Driven Bias Correction and Defect Diagnosis Model for In-Service Vehicle Acceleration Measurements.

Track quality instruments use low-cost accelerometers placed on or attached to the floors of operating trains, and these instruments collect substantial amounts of data over short inspection periods. The measurements collected by the instruments are the main data source for track irregularity evaluation. However, considerable measurement bias exists in the vertical and lateral vibration data obtained from such instruments. False positive track vibration defects detected by track quality instruments occur frequently. This results in considerable time and effort being expended needlessly because maintenance workers have to visit the railway track sites to check and review the track vibration defects. Therefore, we propose a model for data-driven bias correction and defect diagnosis for in-service vehicle acceleration measurements based on track degradation characteristics. Substantial amounts of historical track measurement data from different inspection methods were mined extensively to eliminate the false positive detection of track vibration defects and diagnose the causes of track vibration defects. Actual measurement data from the Lanxin Railway were used to validate our proposed model. The success rate achieved in identifying false positive track vibration defects was 84.1%, and that in track vibration defect diagnosis was 75.8%. These high success rates suggest that the proposed model can be of practical use in improving railway track maintenance management.

Modelling the habitat preferences of the swan mussel (Anodonta cygnea) using data-driven model.

The Anzali wetland (located in northern Iran) and many parts of its catchment are considered important habitats for the swan mussel (Anodonta cygnea). The habitat of this native bioindicator mussel is being threatened in many locations of the catchment due to various anthropogenic activities. The present study aimed to apply a classification tree model (J48 algorithm) to predict the habitat preferences of A. cygnea in 12 sampling sites based on various water quality and physical-habitat variables. The species was present in 50% of sampling sites, while it was absent in the remaining of the sampling sites. In total, 144 samples of A. cygnea (72 presence and 72 absence instances) were monthly measured together with the abiotic variables during 1-year study period (2017-2018). For the CT model, two-thirds of datasets (96 instances) served as a training and the remainder was employed for the validation set (48 instances). Among 25 environmental variables introduced to the model (with pruning confidence factor = 0.10, threefold cross-validation and 5 times randomization effort), the validity of 6 variables was confirmed by the model in all three subsets. Water salinity, flow velocity, water depth and water turbidity were jointly predicted by the model in three subsets. The model predicted that the absence of A. cygnea might be associated with increasing flow velocity, total phosphate and water turbidity. In contrast, the presence of A. cygnea might be related to decreased water depth and increased calcium concentration. The model also confirmed that all predicted variables for the species might be completely dependent on the water salinity. According to the chi-square test (x2 = 26.53, p < 0.05), the habitat condition of A. cygnea is influenced by significant variations in the spatio-temporal patterns.