Capability Of Machine Learning Models Research Articles

Saffron In Vitro Propagation: An Innovative Method by Temporary Immersion System (TIS), Integrated with Machine Learning Analysis

The propagation of Crocus sativus L. relies exclusively on corm multiplication. As underground storage organs, corms are susceptible to a wide range of pathogens, environmental stresses, and diseases, making traditional propagation methods often ineffective with the loss of valuable material. In vitro propagation offers an alternative for the saffron culture under controlled conditions. In particular, the innovative application of the Temporary Immersion System (TIS) represents a technological advancement for enhancing biomass production with a reduction in operational costs. The current study utilized the Plantform™ bioreactor to propagate in vitro saffron corms from the ‘Abruzzo’ region (Italy), integrating machine learning models to assess its performance. The evaluation of saffron explants after 30, 60, and 90 days of culture showed a marked improvement in growth and microcorm production compared to conventional in vitro culture on semisolid medium, supported by the machine learning analysis. Indeed, the Random Forest algorithm revealed a predictive accuracy with an R2 value of 0.81 for microcorm number, showcasing the capability of machine learning models to forecast propagation outcomes effectively. These results confirm that applying TIS in saffron culture could lead to economically viable, large biomass production within a controlled environment, irrespective of seasonality. This study represents the first endeavor to use TIS technology to enhance the in vitro propagation of saffron in conjunction with machine learning, suggesting an innovative approach for cultivating high-value crops like saffron.

Assessment of rainfall-derived inflow and infiltration in sewer systems with machine learning approaches.

Rainfall-derived inflow/infiltration (RDII) modelling during heavy rainfall events is essential for sewer flow management. In this study, two machine learning algorithms, random forest (RF) and long short-term memory (LSTM), were developed for sewer flow prediction and RDII estimation based on field monitoring data. The study implemented feature engineering for extracting physically significant features in sewer flow modelling and investigated the importance of the relevant features. The results from two case studies indicated the superior capability of machine learning models in RDII estimation in the combined and separated sewer systems, and LSTM model outperformed the two models. Compared to traditional methods, machine learning models were capable of simulating the temporal variation in RDII processes and improved prediction accuracy for peak flows and RDII volumes in storm events.

Few-Shot Synthetic Online Transfer Learning for Cross-Site Neurological Disease Diagnosis

Cross-site datasets expand the data size and could improve the disease diagnosis capabilities of machine learning models. However, differences in data distribution between different sites can lead to poor model generalizability. Although transfer learning is a mainstream method often used to tackle the issue, most transfer learning studies assume that all the target samples are given in the training procedure, which is not available in clinical applications where the target samples arrive sequentially. Online transfer learning (OTL) aims to accomplish clinical diagnostic tasks by adaptively updating the ensemble model containing source and target classifiers. However, OTL is limited by zero initialization and requires numbers of samples to iterate. This results in the underperformance of OTL in clinical applications where few samples can be obtained. In this article, we propose a new framework named few-shot synthetic OTL (FSOTL) to address this issue. FSOTL uses synthetic data to warm up the model in an online fashion. It not only alleviates the problem of scarcity of samples in the target domain but also enables the model to gain more knowledge. Our experiments show that FSOTL performs more stably and achieves more accurate results with few target samples, thereby offering a promising cross-site online computer-aided diagnosis system for large-scale applications.

Credit Risk Assessment and Fraud Detection in Financial Transactions Using Machine Learning

Credit risk assessment and fraud detection are crucial tasks in the financial industry, vital to preserving financial organizations' legitimacy and sustainability. Traditional methods often fall short in accurately assessing risk and detecting fraudulent activities in a timely manner. In recent years, machine learning has emerged as a powerful tool for enhancing these processes, leveraging great dimensions of transactional statistics and superior algos for making more informed decisions. This research paper explores the usage of ML techniques in credit risk assessment and fraud detection within financial transactions.
 The paper begins with an overview of the importance of accurate risk assessment and fraud detection in financial transactions and introduces the role of machine learning in addressing these challenges. A comprehensive literature review is conducted to analyze existing methodologies, algorithms, and research trends in the field. Data acquisition and preprocessing techniques are discussed, emphasizing the importance of clean and relevant data for model training. Feature engineering strategies are explored to extract meaningful information from financial transaction data and enhance the predictive capabilities of machine learning models.
 Various machine learning algorithms suitable for credit risk assessment and fraud detection are examined, including LR, SVMs, RF, DTs and DNNs. The efficacy of these techniques is evaluated by discussing model metrics for assessment and ensemble approaches for boosting efficiency, with a focus on metrics such as accuracy, precision, recall, and ROC-AUC.
 The paper presents case studies and experimental results illustrating the application of machine learning models in real-world scenarios, highlighting their effectiveness in improving risk assessment and fraud detection processes. Additionally, difficulties such as imbalanced datasets, comprehensibility of the model and adherence to regulations are discussed, along with potential research directions and future trends in the field.
 In conclusion, this research emphasizes the transformative potential of machine learning in credit risk assessment and fraud detection within financial transactions. By leveraging advanced algorithms and data-driven approaches, financial institutions can enhance their decision-making processes, mitigate risks, and safeguard against fraudulent activities, ultimately contributing to a more secure and resilient financial ecosystem.

Leveraging the power of machine learning and data balancing techniques to evaluate stability in smart grids

Modernizing traditional power grids into smart grids represents a significant advancement in improving efficiency and sustainability within the energy sector. However, integrating renewable energy sources and the intricate network of interconnected devices has presented new challenges, especially in grid stability. Anticipating and maintaining grid stability ensures a continuous energy supply and prevents potential disruptions. This research paper explores the smart grid stability issue through machine learning. Two distinct scenarios are examined. In the first scenario, the problem is approached as a binary classification task, where grid states are categorized as stable or unstable, a well-established method for addressing grid stability. In the second scenario, an effort is made to predict continuous values that reflect the degree of grid stability. These continuous values are subsequently mapped to a conditional function, facilitating stability detection based on a predefined threshold. To address data imbalance, two balancing strategies, specifically the Synthetic Minority Over-sampling Technique (SMOTE) and K-means SMOTE, are employed in each scenario. The analysis covers a range of regressors and classifiers, focusing on tree-based models, boosting algorithms, ensemble learners, and Multilayer Perceptrons (MLP). The investigation results demonstrate that combining the K-means SMOTE strategy with the CatBoost Classifier in the first scenario achieves the highest accuracy, reaching an impressive 99.6%, and an exceptional ability to detect grid instability, achieving 100% accuracy. Furthermore, the results of the second scenario are promising, with accuracy rates reaching 99.4%, thanks to the utilization of a stacking ensemble learner. Notably, this study introduces Explainable AI, a groundbreaking initiative in this field, to delve into the inner workings of the proposed models and enhance their transparency. This pioneering effort marks a significant step forward in understanding and optimizing the predictive capabilities of machine learning models applied to smart grid stability.

Which molecules can challenge density-functional tight-binding methods in evaluating the energies of conformers? investigation with machine-learning toolset

Large organic molecules and biomolecules can adopt multiple conformations, with the occurrences determined by their relative energies. Identifying the energetically most favorable conformations is crucial, especially when interpreting spectroscopic experiments conducted under cryogenic conditions. When the effects of irregular surrounding medium, such as noble gas matrices, on the vibrational properties of molecules become important, semi-empirical (SE) quantum-chemical methods are often employed for computational simulations. Although SE methods are computationally more efficient than first-principle quantum-chemical methods, they can be inaccurate in determining the energies of conformers in some molecules while displaying good accuracy in others. In this study, we employ a combination of advanced machine learning techniques, such as graph neural networks, to identify molecules with the highest errors in the relative energies of conformers computed by the semi-empirical tight-binding method GFN1-xTB. The performance of three different machine learning models is assessed by comparing their predicted errors with the actual errors in conformer energies obtained via the GFN1-xTB method. We further applied the ensemble machine-learning model to a larger collection of molecules from the ChEMBL database and identified a set of molecules as being challenging for the GFN1-xTB method. These molecules hold potential for further improvement of the GFN1-xTB method, showcasing the capability of machine learning models in identifying molecules that can challenge its physical model.

Prediction of chemical reaction yields with large-scale multi-view pre-training

Developing machine learning models with high generalization capability for predicting chemical reaction yields is of significant interest and importance. The efficacy of such models depends heavily on the representation of chemical reactions, which has commonly been learned from SMILES or graphs of molecules using deep neural networks. However, the progression of chemical reactions is inherently determined by the molecular 3D geometric properties, which have been recently highlighted as crucial features in accurately predicting molecular properties and chemical reactions. Additionally, large-scale pre-training has been shown to be essential in enhancing the generalization capability of complex deep learning models. Based on these considerations, we propose the Reaction Multi-View Pre-training (ReaMVP) framework, which leverages self-supervised learning techniques and a two-stage pre-training strategy to predict chemical reaction yields. By incorporating multi-view learning with 3D geometric information, ReaMVP achieves state-of-the-art performance on two benchmark datasets. Notably, the experimental results indicate that ReaMVP has a significant advantage in predicting out-of-sample data, suggesting an enhanced generalization ability to predict new reactions. Scientific Contribution: This study presents the ReaMVP framework, which improves the generalization capability of machine learning models for predicting chemical reaction yields. By integrating sequential and geometric views and leveraging self-supervised learning techniques with a two-stage pre-training strategy, ReaMVP achieves state-of-the-art performance on benchmark datasets. The framework demonstrates superior predictive ability for out-of-sample data and enhances the prediction of new reactions.

Modeling of the biosorption of Fe(III) ions by olive‐stone activated carbon

This study deals with the optimization of Fe(III) ion removal using activated carbon from olive stone waste using advanced machine learning models. The main objective is to evaluate and compare the performance of machine learning models, specifically multilayer perceptron artificial neural network (MLP‐ANN), general regression artificial neural network (GR‐ANN), radial basis function artificial neural network (RBF‐ANN), and particle swarm optimization artificial neural network (PSO‐ANN) in predicting Fe(III) removal efficiency. Experimental data on adsorption parameters were used to train and test the models. Techniques such as tuning hidden layer neurons, optimizing propagation values, and using a Taguchi approach PSO algorithm were applied to improve the models. For the MLP‐ANN model, the optimal configuration contains 13 neurons in the hidden layer. Concerning the parameters involved in the PSO‐ANN model, the coefficient C2 and the particle have the main effect on the reduction of the error. Their contributions are respectively 49% and 19%. The PSO‐ANN model showed superior performance with the highest regression coefficient (0.9997) and remarkable prediction accuracy, surpassing other models such as MLP‐ANN and GR‐ANN. This research suggests that innovative optimization techniques, particularly using PSO algorithms, significantly enhance the predictive capabilities of machine learning models in complex adsorption processes, contributing to more accurate Fe(III) removal models.

Comparisons of Different Machine Learning-Based Rainfall–Runoff Simulations under Changing Environments

Climate change and human activities have a great impact on the environment and have challenged the assumption of the stability of the hydrological time series and the consistency of the observed data. In order to investigate the applicability of machine learning (ML)-based rainfall–runoff (RR) simulation methods under a changing environment scenario, several ML-based RR simulation models implemented in novel continuous and non-real-time correction manners were constructed. The proposed models incorporated categorical boosting (CatBoost), a multi-hidden-layer BP neural network (MBP), and a long short-term memory neural network (LSTM) as the input–output simulators. This study focused on the Dongwan catchment of the Yiluo River Basin to carry out daily RR simulations for the purpose of verifying the model’s applicability. Model performances were evaluated based on statistical indicators such as the deterministic coefficient, peak flow error, and runoff depth error. The research findings indicated that (1) ML-based RR simulation by using a consistency-disrupted dataset exhibited significant bias. During the validation phase for the three models, the R2 index decreased to around 0.6, and the peak flow error increased to over 20%. (2) Identifying data consistency transition points through data analysis and conducting staged RR simulations before and after the transition point can improve simulation accuracy. The R2 values for all three models during both the baseline and change periods were above 0.85, with peak flow and runoff depth errors of less than 20%. Among them, the CatBoost model demonstrated superior phased simulation accuracy and smoother simulation processes and closely matched the measured runoff processes across high, medium, and low water levels, with daily runoff simulation results surpassing those of the BP neural network and LSTM models. (3) When simulating the entire dataset without staged treatment, it is impossible to achieve good simulation results by adopting uniform extraction of the training samples. Under this scenario, the MBP exhibited the strongest generalization capability, highest prediction accuracy, better algorithm stability, and superior simulation accuracy compared to the CatBoost and LSTM simulators. This study offers new ideas and methods for enhancing the runoff simulation capabilities of machine learning models in changing environments.

Deep convolutional neural networks with Bee Collecting Pollen Algorithm (BCPA)-based landslide data balancing and spatial prediction

Predicting the landslide-prone area is critical for various applications, including emergency response, land planning, and disaster mitigation. There needs to be a thorough landslide inventory in current studies and appropriate sampling uncertainty issues. Landslide risk mapping has expanded significantly as machine learning techniques have developed. However, one of the primary issues in Landslide Prediction is data imbalance (DI). This is problematic since it is challenging or expensive to generate an accurate inventory map of landslides based on previous data. This study proposes a novel landslide prediction method using Generative Adversarial Networks (GAN) for generating the synthetic data, Synthetic Minority Oversampling Technique (SMOTE) for overcoming the data imbalance problem, and Bee Collecting Pollen Algorithm (BCPA) for feature extraction. Combining 184 landslides and ten criteria, including topographic wetness index (TWI), aspect, distance from the road, total curvature, sediment transport index (STI), height, slope, stream, lithology, and slope length, a geographical database was produced. The data was generated using GAN, a Deep Convolutional Neural Network (DCNN) technique to populate the dataset. The proposed DCNN-BCPA approach findings were merged with current machine learning methods such as Random Forests (RF), Artificial Neural Networks (ANN), k-Nearest Neighbours (k-NN), Decision Trees (DT), Support Vector Machine (SVM), logistic regression (LR). The model’s accuracy, precision, recall, f-score, and RMSE were measured using the following metrics: 92.675%, 96.298%, 90.536%, 96.637%, and 45.623%. This study suggests that harmonizing landslide data may have a substantial impact on the predictive capabilities of machine learning models.

The Traffic Congestion Prediction Using Machine Learning

Around the world, traffic congestion is a major issue that affects everyday commutes, economic activity, and the sustainability of the environment. The goal of this research project is to use machine learning approaches to address the problem of traffic congestion. Through the utilization of an extensive dataset that encompasses several spatiotemporal parameters, including date, time, weather, and holiday indicators, the research creates predictive models that enable precise forecasting of traffic congestion levels. In order to determine how well machine learning predicts congestion dynamics, the research uses multiple regression methods, such as MLP Regressor, Stacking Regressor, and SVR. Historical traffic volume data is used to train and assess the models, making it possible to identify the main variables impacting patterns of congestion. The findings demonstrate how weather, time of day, and holidays all have a major impact on traffic congestion. Transportation authorities and urban planners may enhance overall urban mobility, optimise infrastructure utilisation, and regulate traffic flows proactively by leveraging the predictive capabilities of machine learning models. This work advances intelligent transport systems by offering a strong framework for anticipating and controlling traffic congestion. The results highlight the significance of utilising data-driven methodologies to tackle intricate urban issues, ultimately cultivating more effective and sustainable urban settings.

Applying machine learning to soil analysis for accurate farming

A crucial component of agriculture is soil. Soil analysis is critical for optimizing agricultural practices and ensuring sustainable crop production. Traditional methods are often time-consuming and labor intensive, limiting their scalability and real-time applicability. The application of machine learning techniques in soil nutrient analysis has emerged as a promising solution. There is a lot of complicated soil data, but algorithms that use machine learning can handle it all, enabling accurate prediction and assessment of soil nutrient content. Integration with remote sensing technologies enhances the capabilities of soil nutrient analysis, allowing for rapid assessment at different scales. Machine learning facilitates personalized recommendations for fertilizer application, irrigation strategies, and soil amendments, tailored to the specific needs of learning and adaptive individual fields or crops. The continuous capabilities of machine learning models ensure up-to-date nutrient management recommendations. Challenges include the availability of representative interpretability of models. Nevertheless, the integration of machine learning in soil nutrient analysis offers improved resource utilization, enhanced crop productivity, and sustainable soil management practices. Ongoing research and collaboration with domain experts will further advance the application of machine learning in this field.

Exploring Feature Engineering Strategies for Improving Predictive Models in Data Science

A crucial step in the data science pipeline, feature engineering has a big impact on how well predictive models function. This study explores several feature engineering techniques and how they affect the robustness and accuracy of models. In order to extract useful information from unprocessed data and improve the prediction capability of machine learning models, we study a variety of techniques, from straightforward transformations to cutting-edge approaches. The study starts by investigating basic methods including data scaling, one-hot encoding, and handling missing values. Then, we go on to more complex techniques like feature selection, dimensionality reduction, and interaction term creation. We also explore the possibilities for domain-specific feature engineering, which entails designing features specifically for the issue domain and utilising additional data sources to expand the feature space. We run extensive experiments on numerous datasets including different sectors, such as healthcare, finance, and natural language processing, in order to evaluate the efficacy of these methodologies. We evaluate model performance using metrics like recall, accuracy, precision, and F1-score to get a comprehensive picture of how feature engineering affects various predictive tasks. This study also assesses the computational expense related to each feature engineering technique, taking scalability and efficiency in practical applications into account. To assist practitioners in making wise choices during feature engineering, we address the trade-offs between model complexity and performance enhancements. Our results highlight the importance of feature engineering in data science and demonstrate how it may significantly improve prediction models in a variety of fields. This study is a useful tool for data scientists because it emphasises the significance of careful feature engineering as a foundation for creating reliable and accurate prediction models.

Adaptive Substation Infrastructure for Bangladesh: Harnessing Artificial Intelligence and Machine Learning for Enhanced Performance and Grid Resilience

This groundbreaking research endeavors to revolutionize substation engineering in Bangladesh by introducing an innovative paradigm that integrates artificial intelligence (AI) and machine learning (ML) methodologies. In response to the dynamic and challenging operational environment, this study focuses on the development of an adaptive substation infrastructure capable of intelligently responding to fluctuating energy demands, environmental stresses, and emerging grid complexities. Through the application of advanced AI algorithms, the research addresses real-time fault detection, predictive maintenance, and comprehensive condition monitoring within the substation framework. Harnessing the capabilities of ML models, the proposed infrastructure aims to optimize energy flow, enhance grid resilience, and mitigate potential failures by autonomously adapting to evolving operational scenarios. By combining cutting-edge technology with the unique challenges of the Bangladeshi power landscape, this research not only aims to advance the field of substation engineering but also holds the promise of significantly contributing to the sustainable development of the nation's power infrastructure. The findings are anticipated to guide the design and implementation of intelligent substation systems, ushering in a new era of efficiency, reliability, and adaptability in the Bangladesh power grid.

A machine learning-driven framework for the property prediction and generative design of multiple principal element alloys

Multi-principal element alloys (MPEAs), inclusive of so-called high entropy alloys (HEAs), represent an innovative class of metallic materials that reveal unique properties and potentially broad applicability. However, the compositional complexity of MPEAs presents challenges in discerning physical mechanisms that control properties, and in harnessing such mechanisms to drive the design of new alloys. An ability to design metallic alloys that possess user-defined requisite properties has emerged as a critical area of interest within the field of materials science and engineering. This research illustrates how the integration of data science, machine learning (ML), and generative design strategies can evolve alloy design to a predictive, data-oriented approach. A comprehensive workflow utilising machine learning for feature analysis, property prediction, and generative design of novel MPEA generation is proposed. This workflow facilitates the examination and comparison of the predictive capabilities of ML models in determining the mechanical properties of MPEAs, providing insights into the influence of different design parameters. Additionally, by integrating a generative adversarial model, the prediction of novel MPEAs and anticipate their mechanical behaviours is revealed.

Nuclear Magnetic Resonance Chemical Shift As Highly Explainable Chemical Structure Fingerprints for Anion Exchange Membrane Polymers

Recent shift towards clean energy increased the demand for both fuel cells (used as clean power generator) and water electrolyzer (used as hydrogen supply) significantly. Anion exchange membrane (AEM) serves as a core component for these devices, though its low anion conductivity and durability inhibit their potential for commercialization. Many research and development (R&D) have been done seeking for improvements in AEM1, but current empirical-centric method consumes significant amount of resources, such as cost, labor, and time. To reduce resource consumptions, implementing materials informatics (MI) that allows high-speed screening of materials through a pre-trained AEM polymer machine learning (ML) model is important. However, AEMs are made up of polymers whose chemical structures are complex and hard to represent in a machine-understandable form. Fingerprints are often used to represent chemical structures in numerical forms generated through algorithm2. Majority of these fingerprints are designed for small molecules, not polymers, but they are usually unintuitive and difficult to understand due to their topological nature. In contrast, nuclear magnetic resonance (NMR) chemical shift have long been used in chemistry to identify the chemical structure of a particular sample3. In this study, we aim to utilize the high-resolution nature of NMR chemical shift to identify structural formula as chemical structure fingerprint for ML model, such that a polymer-suited and highly explainable fingerprint can be developed.First, an AEM database containing structural and experimental condition information was built using data extracted from 62 papers. Experimental conditions included were anion conductivity measuring temperature, alkaline stability test measuring condition (test temperature, length of test, and concentration of alkaline solution). Then, the 13C NMR chemical shift for the chemical structure contained in structural information was calculated using ChemDraw. The obtained chemical shifts were converted to numerical strings and is named as “NMR fingerprint”. A new AEM database containing both structural (molar ratio of each building blocks and NMR fingerprints) and experimental condition was used as the training database for ML models. Target variable was set to anion conductivity, and the rest were explanatory variables. ML model used was XGBoost. Cross-validation was used to evaluate the capability of ML models to predict anion conductivity of novel AEM polymers. Prediction logic was analyzed using Shapley additive explanations (SHAP) value.The database built contains data from 62 AEM papers, with 2,197 anion conductivity data points. Each AEM chemical structures present in the database was converted to NMR fingerprints using NMR chemical shifts, obtaining around 2,000 NMR fingerprints for each AEM polymer unit. Together with the experimental conditions and structural information included in the database, the data were used as train-validation dataset for XGBoost. The coefficient of determination (R2) obtained for cross-validated model was 0.9235, implying that the model learnt and determined the relationship between anion conductivity and AEM polymer structure with high accuracy, with the aid of experimental conditions. Then, the prediction logic of the ML model was explored using SHAP values, which are values computed from coalitional game theory, and is used to increase transparency and interpretability of ML models. Analyzing the plot of SHAP values for top 20 important variables used in XGBoost showed that measuring temperature for anion conductivity ranked highest, which is in coherence to the well-known behavior of AEM polymers. Besides, non-experimental condition variables such as 29.8_A ranked into the top 3 important variables. 29.8_A is the chemical shift for alkyl groups attached in between two imidazolium group of AEM polymer, suggesting that the presence or absence of more than one imidazolium group per side chain is important to determine the anion conductivity of an AEM polymer. SHAP values for 29.8_A show that higher feature value (pink color) gives higher impact (positive region of x-axis) to the target variable, inferring that having alkyl groups between imidazolium groups give beneficial effect to anion conductivity. Such ability to explain the prediction logic of ML model shows that using NMR chemical shifts as fingerprints for AEM polymer structures provide intuitive, human-understandable ML prediction logic explanation. Together with the high cross-validation accuracy, NMR chemical shifts hold the potential to not only be a gold standard in expressing polymer structures in machine-understandable form, but also to strongly push the adoption of ML in the AEM polymer field, creating a paradigm shift for AEM R&D.Reference S. Gottesfeld et al., J. Power Sources 2018, 375, 170-184.J. Bajorath, J. Chem. Inf. Comput. Sci. 2001, 41, 2, 233–245.P. Jezzard et al., Adv. Mater. 1992, 4, 2, 82-90. Figure 1

Ensemble machine learning model trained on a new synthesized dataset generalizes well for stress prediction using wearable devices

Introduction:Advances in wearable sensor technology have enabled the collection of biomarkers that may correlate with levels of elevated stress. While significant research has been done in this domain, specifically in using machine learning to detect elevated levels of stress, the challenge of producing a machine learning model capable of generalizing well for use on new, unseen data remain. Acute stress response has both subjective, psychological and objectively measurable, biological components that can be expressed differently from person to person, further complicating the development of a generic stress measurement model. Another challenge is the lack of large, publicly available datasets labeled for stress response that can be used to develop robust machine learning models. In this paper, we first investigate the generalization ability of models built on datasets containing a small number of subjects, recorded in single study protocols. Next, we propose and evaluate methods combining these datasets into a single, large dataset to study the generalization capability of machine learning models built on larger datasets. Finally, we propose and evaluate the use of ensemble techniques by combining gradient boosting with an artificial neural network to measure predictive power on new, unseen data. In favor of reproducible research and to assist the community advance the field, we make all our experimental data and code publicly available through Github at https://github.com/xalentis/Stress. This paper’s in-depth study of machine learning model generalization for stress detection provides an important foundation for the further study of stress response measurement using sensor biomarkers, recorded with wearable technologies. Methods:Sensor biomarker data from six public datasets were utilized in this study. Exploratory data analysis was performed to understand the physiological variance between study subjects, and the complexity it introduces in building machine learning models capable of detecting elevated levels of stress on new, unseen data. To test model generalization, we developed a gradient boosting model trained on one dataset (SWELL), and tested its predictive power on two datasets previously used in other studies (WESAD, NEURO). Next, we merged four small datasets, i.e. (SWELL, NEURO, WESAD, UBFC-Phys), to provide a combined total of 99 subjects, and applied feature engineering to generate additional features utilizing statistical summaries, with sliding windows of 25 s. We name this large dataset, StressData. In addition, we utilized random sampling on StressData combined with another dataset (EXAM) to build a larger training dataset consisting of 200 synthesized subjects, which we name SynthesizedStressData. Finally, we developed an ensemble model that combines our gradient boosting model with an artificial neural network, and tested it using Leave-One-Subject-Out (LOSO) validation, and on two additional, unseen publicly available stress biomarker datasets (WESAD and Toadstool). Results:Our results show that previous models built on datasets containing a small number (<50) of subjects, recorded in single study protocols, cannot generalize well to new, unseen datasets. Our presented methodology for generating a large, synthesized training dataset by utilizing random sampling to construct scenarios closely aligned with experimental conditions demonstrate significant benefits. When combined with feature-engineering and ensemble learning, our method delivers a robust stress measurement system capable of achieving 85% predictive accuracy on new, unseen validation data, achieving a 25% performance improvement over single models trained on small datasets. The resulting model can be used as both a classification or regression predictor for estimating the level of perceived stress, when applied on specific sensor biomarkers recorded using a wearable device, while further allowing researchers to construct large, varied datasets for training machine learning models that closely emulate their exact experimental conditions. Conclusion:Models trained on small, single study protocol datasets do not generalize well for use on new, unseen data and lack statistical power. Machine learning models trained on a dataset containing a larger number of varied study subjects capture physiological variance better, resulting in more robust stress detection. Feature-engineering assists in capturing these physiological variance, and this is further improved by utilizing ensemble techniques by combining the predictive power of different machine learning models, each capable of learning unique signals contained within the data. While there is a general lack of large, labeled public datasets that can be utilized for training machine learning models capable of accurately measuring levels of acute stress, random sampling techniques can successfully be applied to construct larger, varied datasets from these smaller sample datasets, for building robust machine learning models.

Automated Machine Learning-Based Diagnosis and Molecular Characterization of Acute Leukemias using Flow Cytometry Data

Abstract Current flow cytometric analysis of blood and bone marrow samples for the diagnosis of acute leukemias relies heavily on manual intervention in both the processing and analysis steps, introducing significant subjectivity into the resulting diagnosis and increasing diagnostic turn-around time. Additionally, concurrent molecular characterization of these samples via cytogenetics and targeted sequencing panels can take multiple days, thereby delaying patient diagnosis and treatment. Attention-based multi-instance learning models are machine learning models that can make accurate predictions and generate interpretable insights regarding the classification of a sample from multiple events/cells; while these models have been developed for anatomic pathology applications, they have yet to be applied to flow cytometry data. By utilizing 1,820 flow cytometry samples from 2019-2022 at Brigham and Women’s Hospital, we developed attention-based multi-instance machine learning models for automated diagnosis of acute leukemia, including differentiation of acute myeloid leukemia (AML) from B-lymphoblastic leukemia/lymphoma (B-ALL). Additionally, using concurrent cytogenetic and targeted sequencing data from 674 acute leukemia samples, machine learning models for prediction of molecular aberrancies from flow cytometry data were developed. Machine learning models were created using the TabNet deep learning architecture and VIME self-supervised training algorithm, which are state-of-the-art approaches towards machine learning from tabular data including flow cytometry data. Attention-based multi-instance models demonstrated strong performance for the automated diagnosis of acute leukemia versus non-leukemia samples (AUROC 0.869), as well as the separation of AML from B-ALL samples (AUROC 0.971). These models also accurately predicted cytogenetic aberrancies among AML samples, including t(15;17);PML::RARA (AUROC 1.00), as well as mutations including NPM1 (AUROC 0.725). These models do not require any manual intervention including compensation or gating, and additionally provide quantitative scores for the relative importance of different flow cytometry events and markers for the diagnosis of a particular sample. These importance scores can be integrated into flow cytometry analysis software for visualization and interpretation by hematopathologists. In this study, we have demonstrated the capability of machine learning models to provide automated diagnoses of acute leukemia, as well as accurately predict cytogenetic and molecular aberrancies in blood and bone marrow samples using flow cytometry data. This automated workflow can significantly decrease diagnostic turn-around time and ultimately improve patient outcomes.

A comparative study of machine learning regression models for predicting construction duration

ABSTRACT Over the past few decades, the construction industry has been suffering from project delays. It remains a recognized challenge to accurately predict the actual project’s progress, despite the vast amount of field data and scheduling methods available. In this paper, four different machine learning (ML) models including K-nearest neighbor (KNN), support vector regression (SVR), gradient boosting trees (GBT), and artificial neural network (ANN) were used for forecasting the construction duration of work areas. Field data were collected, including progress, workload, labor, weather, planned days, and location, as inputs and construction days as output for each work area to develop the models. A simple and widely used baseline model was used to confirm the accuracy and generalization capability of ML models. The comparison of the results revealed that ANN and GBT models produced significantly superior results. Additionally, the GBT model is substantially more computationally efficient. In conclusion, the GBT model could be successfully and effectively applied to improve the prediction of construction duration. The proposed ML models could be utilized as a decision support tool for construction project managers to adjust construction site resource allocation.

ML-Based Teaching Systems: A Conceptual Framework

As the shortage of skilled workers continues to be a pressing issue, exacerbated by demographic change, it is becoming a critical challenge for organizations to preserve the knowledge of retiring experts and pass it on to novices. While this knowledge transfer has traditionally occurred through personal interaction, it lacks scalability and requires significant resources and time. IT-based teaching systems have addressed this scalability issue, but their development is still tedious and time-consuming. In this work, we investigate the potential of machine learning (ML) models to facilitate knowledge transfer in an organizational context, leading to more cost-effective IT-based teaching systems. Through a systematic literature review, we examine key concepts, themes, and dimensions to understand better and design ML-based teaching systems. To do so, we capture and consolidate the capabilities of ML models in IT-based teaching systems, inductively analyze relevant concepts in this context, and determine their interrelationships. We present our findings in the form of a review of the key concepts, themes, and dimensions to understand and inform on ML-based teaching systems. Building on these results, our work contributes to research on computer-supported cooperative work by conceptualizing how ML-based teaching systems can preserve expert knowledge and facilitate its transfer from SMEs to human novices. In this way, we shed light on this emerging subfield of human-computer interaction and serve to build an interdisciplinary research agenda.