Integrating local climate considerations into dormitory energy optimization: An explainable machine learning and multi-objective design approach

The Open University UK has been designing, implementing, and evaluating its OU Learning Design Initiative (OULDI) for over 15 years. Since 2015 a range of learning analytics (LA) studies and practical intervention studies have shown that these learning design (LD) decisions made by educators substantially influence what, how and when students are learning. However, applying and translating OULDI and LA in other institutions and borders is not a mere copy-paste job, and over the years OULDI been adopted and refined in a range of institutions. In this contribution we aim to reflect on how the principles and practical approaches of OULDI and LA need to be adjusted to fit other institutions and approaches. In several EU projects working with 64 practitioners from ten institutions from nine countries, University of Zagreb has critically reflected on the OULDI approach as well as other LD approaches, and altered several substantial elements in a newly designed Balanced Design Planning (BDP) approach. The lessons learned of applying and translating LD and LA approaches could help to inform educators how to use existing LD and LA approaches and adapt them to their institutional needs.

Objective: The number of patients requiring prolonged mechanical ventilation (PMV) is increasing worldwide, but the weaning outcome prediction model in these patients is still lacking. We hence aimed to develop an explainable machine learning (ML) model to predict successful weaning in patients requiring PMV using a real-world dataset.Methods: This retrospective study used the electronic medical records of patients admitted to a 12-bed respiratory care center in central Taiwan between 2013 and 2018. We used three ML models, namely, extreme gradient boosting (XGBoost), random forest (RF), and logistic regression (LR), to establish the prediction model. We further illustrated the feature importance categorized by clinical domains and provided visualized interpretation by using SHapley Additive exPlanations (SHAP) as well as local interpretable model-agnostic explanations (LIME).Results: The dataset contained data of 963 patients requiring PMV, and 56.0% (539/963) of them were successfully weaned from mechanical ventilation. The XGBoost model (area under the curve [AUC]: 0.908; 95% confidence interval [CI] 0.864–0.943) and RF model (AUC: 0.888; 95% CI 0.844–0.934) outperformed the LR model (AUC: 0.762; 95% CI 0.687–0.830) in predicting successful weaning in patients requiring PMV. To give the physician an intuitive understanding of the model, we stratified the feature importance by clinical domains. The cumulative feature importance in the ventilation domain, fluid domain, physiology domain, and laboratory data domain was 0.310, 0.201, 0.265, and 0.182, respectively. We further used the SHAP plot and partial dependence plot to illustrate associations between features and the weaning outcome at the feature level. Moreover, we used LIME plots to illustrate the prediction model at the individual level. Additionally, we addressed the weekly performance of the three ML models and found that the accuracy of XGBoost/RF was ~0.7 between weeks 4 and week 7 and slightly declined to 0.6 on weeks 8 and 9.Conclusion: We used an ML approach, mainly XGBoost, SHAP plot, and LIME plot to establish an explainable weaning prediction ML model in patients requiring PMV. We believe these approaches should largely mitigate the concern of the black-box issue of artificial intelligence, and future studies are warranted for the landing of the proposed model.

Multi-objective design approach of passive filters for single-phase distributed energy grid integration systems using particle swarm optimization

Purpose: Maternal mortality remained to persist in many developing countries. India being the most populous developing country has several cultural differences and beliefs on health systems and may have disparities in receiving proper maternal health care. High data availability and under-utilization of data centric decision making is a key reason for ineffective performance of health systems. Enacting machine learning on such data to aid in sub-divisional policy formulation to address area level problems will eradicate major disparities in recipients of health services. Methods: Glass box machine learning models are trained on the data to obtain importance of features in defining the maternal mortality of a district. Furthermore, black box machine learning models are trained with hyper-parameter tuning and best model is chosen to perform explainable machine learning to generate explanations for each district prediction. A hybrid explainable machine learning approach is proposed on black-box machine learning models where Shapley Additive Explanation and Local Interpretable Model-agnostic Explanations are combined to generate final explanations. Results: There may be several differences even among nearby districts. Health Management Information System data is analyzed with help of Machine Learning techniques and Explainable Machine Learning techniques are used on the trained models to evaluate the contributing factors for each district. Conclusion: The factors that are specific to each district can help in formulating region specific health policies that minimize the disparities of progress of preventing maternal mortality over the districts of India. The paper has highlighted the advantages of using explainable machine learning in extracting complicated patterns of the data.

Identifying soil drivers of rice productivity under fly ash and organic amendments using explainable machine learning.

Efficient Milling Quality Prediction with Explainable Machine Learning

In this study, a multi-objective design approach with accelerated methodology was developed for a B-pillar (side door pillar) in which the intrusion velocity was decreased and the crash energy absorbed. The B-pillar material characteristics were determined using a drop tower test to accelerate the design process instead of a vehicle crash test. A finite element simulation of the drop tower test was conducted, and the results obtained from the simulation were confirmed with the test results. The side impact finite element model was simulated according to the Euro NCAP test protocol, and the B-pillar was divided into two sections using the results obtained from the analysis. Tailor rolled blank and Tailor welded blank B-pillar crash simulations were performed, and the results were compared to examine the intrusion velocity. Alternative design solutions for single and multi-material B-pillars were conducted to design an optimum B-pillar structure. The design functions were created using the radial basis function method, and the failure criteria were determined for the analyses. Optimization problems for weight minimization and maximum energy absorption were established and solved using meta-heuristic methods. The approach suggested in this study can be used in accelerated B-pillar designs.

BackgroundWeaning from mechanical ventilation (MV) is an essential issue in critically ill patients, and we used an explainable machine learning (ML) approach to establish an extubation prediction model.MethodsWe enrolled patients who were admitted to intensive care units during 2015–2019 at Taichung Veterans General Hospital, a referral hospital in central Taiwan. We used five ML models, including extreme gradient boosting (XGBoost), categorical boosting (CatBoost), light gradient boosting machine (LightGBM), random forest (RF) and logistic regression (LR), to establish the extubation prediction model, and the feature window as well as prediction window was 48 h and 24 h, respectively. We further employed feature importance, Shapley additive explanations (SHAP) plot, partial dependence plot (PDP) and local interpretable model-agnostic explanations (LIME) for interpretation of the model at the domain, feature, and individual levels.ResultsWe enrolled 5,940 patients and found the accuracy was comparable among XGBoost, LightGBM, CatBoost and RF, with the area under the receiver operating characteristic curve using XGBoost to predict extubation was 0.921. The calibration and decision curve analysis showed well applicability of models. We also used the SHAP summary plot and PDP plot to demonstrate discriminative points of six key features in predicting extubation. Moreover, we employed LIME and SHAP force plots to show predicted probabilities of extubation and the rationale of the prediction at the individual level.ConclusionsWe developed an extubation prediction model with high accuracy and visualised explanations aligned with clinical workflow, and the model may serve as an autonomous screen tool for timely weaning.

This study proposes a multi-objective optimal design approach for water distribution systems, considering mechanical system redundancy under multiple pipe failure. Mechanical redundancy is applied to the system’s hydraulic ability, based on the pressure deficit between the pressure requirements under abnormal conditions. The developed design approach shows the relationships between multiple pipe failure states and system redundancy, for different numbers of pipe-failure conditions (e.g., first, second, third, …, tenth). Furthermore, to consider extreme demand modeling, the threshold of the demand quantity is investigated simultaneously with multiple pipe failure modeling. The design performance is evaluated using the mechanical redundancy deficit under extreme demand conditions. To verify the proposed design approach, an expanded version of the well-known benchmark network is used, configured as an ideal grid-shape, and the multi-objective harmony search algorithm is used as the optimal design approach, considering construction cost and system mechanical redundancy. This optimal design technique could be used to propose a standard for pipe failure, based on factors such as the number of broken pipes, during failure condition analysis for redundancy-based designs of water distribution systems.

This study explores the influence of concrete mix ingredients on the non-steady chloride migration coefficient (Dnssm) using an explainable machine learning (XML) approach that integrates Extreme Gradient Boosting (XGBoost) and Shapley Additive Explanations (SHAP). The dataset, comprising 204 observations from literature, is utilized to train the XGBoost algorithm for predicting Dnssm. The model demonstrates notable performance metrics with (MAE = 1.61 × 10−12 m2/s, RMSE = 2.38 × 10−12 m2/s, and R2 = 0.95) in the training set and (MAE = 2.22 × 10−12 m2/s, RMSE = 3.18 × 10−12 m2/s, and R2 = 0.87) and the test set. The SHAP method provides comprehensive insights into feature importance, offering valuable information about the relationships and dependencies among various features. The top five features identified as significant contributors include coarse aggregate, superplasticizer, concrete age, cement, and water. Visualization of SHAP values through diverse plots proves essential for obtaining a thorough understanding of feature influence. The explainability of the model's results contributes new insights, aiding in the development of optimal and sustainable concrete with enhanced resistance to chloride penetration. Furthermore, the model's explainability fosters trust in its predictions, facilitating seamless integration into real-world applications.

A combined inductive learning and experimental design approach to manufacturing operation planning

Conceptual design is a fundamental phase for developing optimal product configurations. During conceptual design, the degree of freedom in engineering choices can propose optimal solutions in terms of assembly, manufacturing, cost and material selection. Nevertheless, in current industrial practices, each aspect is analysed independently and a guided decision-making approach based on multi-objective criteria is missing. Multi-objective analysis is a way of combining each production aspect with the aim of choosing the best design option. The goal of this research work is to define a multi-objective design approach for the determination of optimal and feasible design options during the conceptual design phase. The approach is based on the concept of functional basis, module heuristics for defining product modules and the theory of multi-criteria decision-making for mathematical assessment of the best design option. The novelty of this approach lies in making the design process, currently based on company know-how and experience, systematic. A complex product (i.e. tool-holder carousel of a computer numerical control machine tool) is the case study used to assess the economic sustainability of different design options and to validate the proposed design workflow in a real manufacturing context. Different product modules have been re-designed and prototyped for comparing the assemblability, manufacturability and cost of the design solutions.

Enhancing efficiency and environmental performance of laser-cutting machine tools: An explainable machine learning approach

Natural fiber-reinforced plastic (NFRP) composites are ecofriendly and biodegradable materials that offer tremendous ecological advantages while preserving unique structures and properties. Studies on using these natural fibers as alternatives to conventional synthetic fibers in fiber-reinforced materials have opened up possibilities for industrial applications, especially for sustainable manufacturing. However, critical issues reside in the machinability of such materials because of their multiscale structure and the randomness of the reinforcing elements distributed within the matrix basis. This paper reports a comprehensive investigation of the effect of microstructure heterogeneity on the resultant behaviors of cutting forces for NFRP machining. A convolutional neural network (CNN) links the microstructural reinforcing fibers and their impacts on changing the cutting forces (with an estimated R-squared value over 90%). Next, a model-agnostic explainable machine learning approach is implemented to decipher this CNN black-box model by discovering the underlying mechanisms of relating the reinforcing elements/fibers’ microstructures. The presented xml approach extracts physical descriptors from the in-process monitoring microscopic images and finds the causality of the fibrous structures’ heterogeneity to the resultant machining forces. The results suggest that, for the heterogeneous fibers, the tightly and evenly bounded fiber elements (i.e., with lower aspect ratio, lower eccentricity, and higher compactness) strengthen the material and thereafter play a significant role in increasing the cutting forces during NFRP machining. Therefore, the presented framework of the explainable machine learning approach opens an opportunity to discover the causality of material microstructures on the resultant process dynamics and accurately predict the cutting behaviors during material removal processes.

BackgroundThis study aimed to develop a higher performance nomogram based on explainable machine learning methods, and to predict the risk of death of stroke patients within 30 days based on clinical characteristics on the first day of intensive care units (ICU) admission.MethodsData relating to stroke patients were extracted from the Medical Information Marketplace of the Intensive Care (MIMIC) IV and III database. The LightGBM machine learning approach together with Shapely additive explanations (termed as explain machine learning, EML) was used to select clinical features and define cut-off points for the selected features. These selected features and cut-off points were then evaluated using the Cox proportional hazards regression model and Kaplan-Meier survival curves. Finally, logistic regression-based nomograms for predicting 30-day mortality of stroke patients were constructed using original variables and variables dichotomized by cut-off points, respectively. The performance of two nomograms were evaluated in overall and individual dimension.ResultsA total of 2982 stroke patients and 64 clinical features were included, and the 30-day mortality rate was 23.6% in the MIMIC-IV datasets. 10 variables (“sofa (sepsis-related organ failure assessment)”, “minimum glucose”, “maximum sodium”, “age”, “mean spo2 (blood oxygen saturation)”, “maximum temperature”, “maximum heart rate”, “minimum bun (blood urea nitrogen)”, “minimum wbc (white blood cells)” and “charlson comorbidity index”) and respective cut-off points were defined from the EML. In the Cox proportional hazards regression model (Cox regression) and Kaplan-Meier survival curves, after grouping stroke patients according to the cut-off point of each variable, patients belonging to the high-risk subgroup were associated with higher 30-day mortality than those in the low-risk subgroup. The evaluation of nomograms found that the EML-based nomogram not only outperformed the conventional nomogram in NIR (net reclassification index), brier score and clinical net benefits in overall dimension, but also significant improved in individual dimension especially for low “maximum temperature” patients.ConclusionsThe 10 selected first-day ICU admission clinical features require greater attention for stroke patients. And the nomogram based on explainable machine learning will have greater clinical application.