Search

Excessive vibration of stay cables in strong winds has been a concern for bridge operators, which impairs the durability of both the cables and the bridge structure. This paper develops a data-driven approach to predict the amplitude of the cable vibration using an ensemble learning model. The model aims to predict cable vibrations in both in-plane and out-of-plane directions, with the wind speed, wind direction, turbulence intensity, and deck acceleration as input variables. Especially, the deck acceleration is included considering the deck-cable interaction and vehicle effects, which significantly improved the accuracy of the prediction. Furthermore, the model is interpreted with local interpretable model-agnostic explanations (LIME) and partial dependence plot (PDP) methods. The former demonstrates the relative importance of input variables on a global scale, and the latter indicates the correlation between individual input variables with the prediction target. The investigation is validated using the data harnessed from structural health monitoring (SHM) of a 1088-m cable-stayed bridge during three typhoon events. The adopted Gradient boosting regression tree (GBRT) model demonstrated better performance than other state-of-the-art machine learning models. The developed approach can provide guidance on preventive maintenance of stay cables to avoid damage due to excessive vibration. [ABSTRACT FROM AUTHOR]

Explainable artificial intelligence (XAI) uses artificial intelligence (AI) tools and techniques to build interpretability in black-box algorithms. XAI methods are classified based on their purpose (pre-model, in-model, and post-model), scope (local or global), and usability (model-agnostic and model-specific). XAI methods and techniques were summarized in this paper with real-life examples of XAI applications. Local interpretable model-agnostic explanations (LIME) and shapley additive explanations (SHAP) methods were applied to the moral dataset to compare the performance outcomes of these two methods. Through this study, it was found that XAI algorithms can be custom-built for enhanced model-specific explanations. There are several limitations to using only one method of XAI and a combination of techniques gives complete insight for all stakeholders. [ABSTRACT FROM AUTHOR]

Laser-induced breakdown spectroscopy (LIBS) technology faces the challenge of redundant or irrelevant features when dealing with high-dimensional data of steel. To enhance the accuracy and interpretability of multivariate classification, this study introduces an innovative hybrid feature selection (FS) method that skillfully combines the filtering characteristics of the select percentile (SP) algorithm with the embedded advantages of the elastic net (EN) algorithm. Under this framework, the support vector machine (SVM) algorithm was applied for classification, demonstrating outstanding performance with an accuracy, precision, and F1 score of 0.9888, 0.9895, and 0.9889 on the test set, respectively. To address the 'black box' nature of the SVM algorithm, this paper further introduces the local interpretable model-agnostic explanations (LIME) method. LIME allows for the visualization of the importance of each variable, thereby enhancing the interpretability and credibility of the model. Overall, the model and methods proposed in this study show significant effectiveness in eliminating redundant or irrelevant features and in precise classification, effectively solving most of the challenges faced by LIBS in steel classification issues. [ABSTRACT FROM AUTHOR]

As the world grapples with pandemics and increasing stress levels among individuals, heart failure (HF) has emerged as a prominent cause of mortality on a global scale. The most effective approach to improving the chances of individuals’ survival is to diagnose this condition at an early stage. Researchers widely utilize supervised feature selection techniques alongside conventional standalone machine learning (ML) algorithms to achieve the goal. However, these approaches may not consistently demonstrate robust performance when applied to data that they have not encountered before, and struggle to discern intricate patterns within the data. Hence, we present a Multi-objective Stacked Enable Hybrid Model (MO-SEHM), that aims to find out the best feature subsets out of numerous different sets, considering multiple objectives. The Stacked Enable Hybrid Model (SEHM) plays the role of classifier and integrates with a multi-objective feature selection method, the Non-dominated Sorting Genetic Algorithm II (NSGA-II). We employed an HF dataset from the Faisalabad Institute of Cardiology (FIOC) and evaluated six ML models, including SEHM with and without NSGA-II for experimental purposes. The Pareto front (PF) demonstrates that our introduced MO-SEHM surpasses the other models, obtaining 94.87% accuracy with the nine relevant features. Finally, we have applied Local Interpretable Model-agnostic Explanations (LIME) with MO-SEHM to explain the reasons for individual outcomes, which makes our model transparent to the patients and stakeholders.

Purpose – This study aims to examine foreign direct investment (FDI) factors and develops a rational framework for FDI inflow in Western European countries such as France, Germany, the Netherlands, Switzerland, Belgium and Austria. Design/methodology/approach – Data for this study were collected from the World development indicators (WDI) database from 1995 to 2018. Factors such as economic growth, pollution, trade, domestic capital investment, gross value-added and the financial stability of the country that influence FDI decisions were selected through empirical literature. A framework was developed using interpretable machine learning (IML), decision trees and three-stage least squares simultaneous equation methods for FDI inflow in Western Europe. Findings – The findings of this study show that there is a difference between the most important and trusted factors for FDI inflow. Additionally, this study shows that machine learning (ML) models can perform better than conventional linear regression models. Research limitations/implications – This research has several limitations. Ideally, classification accuracies should be higher, and the current scope of this research is limited to examining the performance of FDI determinants within Western Europe. Practical implications – Through this framework, the national government can understand how investors make their capital allocation decisions in their country. The framework developed in this study can help policymakers better understand the rationality of FDI inflows. Originality/value – An IML framework has not been developed in prior studies to analyze FDI inflows. Additionally, the author demonstrates the applicability of the IML framework for estimating FDI inflows in Western Europe.

Abstract Objective Sepsis-induced coagulopathy (SIC) is extremely common in individuals with sepsis, significantly associated with poor outcomes. This study attempted to develop an interpretable and generalizable machine learning (ML) model for early predicting the risk of 28-day death in patients with SIC. Methods In this retrospective cohort study, we extracted SIC patients from the Medical Information Mart for Intensive Care III (MIMIC-III), MIMIC-IV, and eICU-CRD database according to Toshiaki Iba's scale. And the overlapping in the MIMIC-IV was excluded for this study. Afterward, only the MIMIC-III cohort was randomly divided into the training set, and the internal validation set according to the ratio of 7:3, while the MIMIC-IV and eICU-CRD databases were considered the external validation sets. The predictive factors for 28-day mortality of SIC patients were determined using recursive feature elimination combined with tenfold cross-validation (RFECV). Then, we constructed models using ML algorithms. Multiple metrics were used for evaluation of performance of the models, including the area under the receiver operating characteristic curve (AUROC), area under the precision recall curve (AUPRC), accuracy, sensitivity, specificity, negative predictive value, positive predictive value, recall, and F1 score. Finally, Shapley Additive Explanations (SHAP), Local Interpretable Model-Agnostic Explanations (LIME) were employed to provide a reasonable interpretation for the prediction results. Results A total of 3280, 2798, and 1668 SIC patients were screened from MIMIC-III, MIMIC-IV, and eICU-CRD databases, respectively. Seventeen features were selected to construct ML prediction models. XGBoost had the best performance in predicting the 28-day mortality of SIC patients, with AUC of 0.828, 0.913 and 0.923, the AUPRC of 0.807, 0.796 and 0.921, the accuracy of 0.785, 0.885 and 0.891, the F1 scores were 0.63, 0.69 and 0.70 in MIMIC-III (internal validation set), MIMIC-IV, and eICU-CRD databases. The importance ranking and SHAP analyses showed that initial SOFA score, red blood cell distribution width (RDW), and age were the top three critical features in the XGBoost model. Conclusions We developed an optimal and explainable ML model to predict the risk of 28-day death of SIC patients 28-day death risk. Compared with conventional scoring systems, the XGBoost model performed better. The model established will have the potential to improve the level of clinical practice for SIC patients. Graphical Abstract

Machine learning-based systems have presented increasing learning performance, in a wide variety of tasks. However, the problem with some state-of-the-art models is their lack of transparency, trustworthiness, and explainability. To address this problem, eXplainable Artificial Intelligence (XAI) appeared. It is a research field that aims to make black-box models more understandable to humans. The research on this topic has increased in recent years, and many methods, such as LIME (Local Interpretable Model-Agnostic Explanations) and SHAP (SHapley Additive exPlanations) have been proposed. Machine learning-based Intrusion Detection Systems (IDS) are one of the many application domains of XAI. However, most of the works about model interpretation focus on other fields, like computer vision, natural language processing, biology, healthcare, etc. This poses a challenge for cybersecurity professionals tasked with analyzing IDS results, thereby impeding their capacity to make informed decisions. In an attempt to address this problem, we have selected two XAI methods, LIME, and SHAP. Using the methods, we have retrieved explanations for the results of a black-box model, part of an IDS solution that performs intrusion detection on IoT devices, increasing its interpretability. In order to validate the explanations, we carried out a perturbation analysis where we tried to obtain a different classification based on the features present in the explanations. With the explanations and the perturbation analysis we were able to draw conclusions about the negative impact of particular features on the model results when present in the input data, making it easier for cybersecurity experts when analyzing the model results and it serves as an aid to the continuous improvement the model. The perturbations also serve as a comparison of performance between LIME and SHAP. To evaluate the degree of interpretability increase, and the explanations provided by each XAI method of the model and directly compare the XAI methods, we have performed a survey analysis.

Objective: Sepsis-induced coagulopathy (SIC) is extremely common in individuals with sepsis, significantly associated with poor outcomes. This study attempted to develop an interpretable and generalizable machine learning (ML) model for early predicting the risk of 28-day death in patients with SIC. Methods: In this retrospective cohort study, we extracted SIC patients from the Medical Information Mart for Intensive Care III (MIMIC-III), MIMIC-IV, and eICU-CRD database according to Toshiaki Iba's scale. And the overlapping in the MIMIC-IV was excluded for this study. Afterward, only the MIMIC-III cohort was randomly divided into the training set, and the internal validation set according to the ratio of 7:3, while the MIMIC-IV and eICU-CRD databases were considered the external validation sets. The predictive factors for 28-day mortality of SIC patients were determined using recursive feature elimination combined with tenfold cross-validation (RFECV). Then, we constructed models using ML algorithms. Multiple metrics were used for evaluation of performance of the models, including the area under the receiver operating characteristic curve (AUROC), area under the precision recall curve (AUPRC), accuracy, sensitivity, specificity, negative predictive value, positive predictive value, recall, and F1 score. Finally, Shapley Additive Explanations (SHAP), Local Interpretable Model-Agnostic Explanations (LIME) were employed to provide a reasonable interpretation for the prediction results. Results: A total of 3280, 2798, and 1668 SIC patients were screened from MIMIC-III, MIMIC-IV, and eICU-CRD databases, respectively. Seventeen features were selected to construct ML prediction models. XGBoost had the best performance in predicting the 28-day mortality of SIC patients, with AUC of 0.828, 0.913 and 0.923, the AUPRC of 0.807, 0.796 and 0.921, the accuracy of 0.785, 0.885 and 0.891, the F1 scores were 0.63, 0.69 and 0.70 in MIMIC-III (internal validation set), MIMIC-IV, and eICU-CRD databases. The importance ranking and SHAP analyses showed that initial SOFA score, red blood cell distribution width (RDW), and age were the top three critical features in the XGBoost model. Conclusions: We developed an optimal and explainable ML model to predict the risk of 28-day death of SIC patients 28-day death risk. Compared with conventional scoring systems, the XGBoost model performed better. The model established will have the potential to improve the level of clinical practice for SIC patients. [ABSTRACT FROM AUTHOR]

Owing to the convenience of online loans, an increasing number of people are borrowing money on online platforms. With the emergence of machine learning technology, predicting loan defaults has become a popular topic. However, machine learning models have a black-box problem that cannot be disregarded. To make the prediction model rules more understandable and thereby increase the user’s faith in the model, an explanatory model must be used. Logistic regression, decision tree, XGBoost, and LightGBM models are employed to predict a loan default. The prediction results show that LightGBM and XGBoost outperform logistic regression and decision tree models in terms of the predictive ability. The area under curve for LightGBM is 0.7213. The accuracies of LightGBM and XGBoost exceed 0.8. The precisions of LightGBM and XGBoost exceed 0.55. Simultaneously, we employed the local interpretable model-agnostic explanations approach to undertake an explainable analysis of the prediction findings. The results show that factors such as the loan term, loan grade, credit rating, and loan amount affect the predicted outcomes.

In the field of environmental art design, there are significant differences in the level of designers. Some designers lack innovation capabilities and technical support, making it difficult to meet complex design needs, affecting the overall design quality and efficiency. There is also a lack of effective interaction and system between some automated design systems and designers. Problems such as poor interpretability lead to low overall design efficiency and lack of innovation. To address these issues, this article studies an environmental art automated design method based on machine learning (ML) in the field of commercial spaces. This study aims to use machine learning technology to build an automated method for environmental art design, improve design efficiency and optimize creative generation through intelligent models, promote the transformation of environmental art design from a traditional model to an innovative model driven by data and supported by algorithms, and achieve efficient and personalized design goals. Firstly, the image data of commercial space design is normalized and combined with labels such as functional attributes and design styles. Generative adversarial networks (GANs) are used to generate design schemes, and the diversity of design schemes is optimized through feature fusion technology. The interpretability of the design is enhanced through ensemble learning and local interpretable model-agnostic explanations (LIMEs) model, enabling designers to understand and optimize the decision-making process of the model. The experimental results show that compared with the traditional manual design, the longest time of the automated design method in design efficiency testing is 18.1 hours, far shorter than that of the manual design method. In design diversity testing, the automated design method is more diverse in terms of style and number of elements. Meanwhile, the automated design scheme overall outperforms manual design in terms of space utilization, and the average score for layout decisions in interpretability testing reaches a maximum of 9.5 points. These data clearly demonstrate the advantages of the automated design method in improving design efficiency and innovation, and it is easy for designers to understand. Methods based on machine learning have significantly improved the efficiency, quality, and innovation of environmental art design, providing intelligent solutions for the design industry. [ABSTRACT FROM AUTHOR]

Lung cancer significantly contributes to global cancer mortality, posing challenges in clinical management. Early detection and accurate prognosis are crucial for improving patient outcomes. This study develops an interpretable stacking ensemble model (SEM) for lung cancer prognosis prediction and identifies key risk factors. Using a Kaggle dataset of 1000 patients with 22 variables, the model classifies prognosis into Low, Medium, and High-risk categories. The bootstrap method was employed for evaluation metrics, while SHAP (Shapley Additive Explanations) and LIME (Local Interpretable Model-agnostic Explanations) assessed model interpretability. Results showed SEM's superior interpretability over traditional models, such as Random Forest, Logistic Regression, Decision Tree, Gradient Boosting Machine, Extreme Gradient Boosting Machine, and Light Gradient Boosting Machine. SEM achieved an accuracy of 98.90 %, precision of 98.70 %, F1 score of 98.85 %, sensitivity of 98.77 %, specificity of 95.45 %, Cohen's kappa value of 94.56 %, and an AUC of 98.10 %. The SEM demonstrated robust performance in lung cancer prognosis, revealing chronic lung cancer and genetic risk as major factors. [Display omitted] • Developed an interpretable stacking ensemble model (SEM) for lung cancer prognosis prediction. • SEM outperformed traditional models with an accuracy of 98.90 %, precision of 98.70, sensitivity of 98.77 %, and AUC of 98.10 %. • Integrated SHAP and LIME techniques to enhance model interpretability and facilitate clinical decision-making. • Identified chronic lung cancer, genetic risk, and other key factors as significant predictors of lung cancer prognosis. [ABSTRACT FROM AUTHOR]

In advancement of interpretable machine learning (IML), this research proposes local interpretable model-agnostic explanations (LIME) as a new visualization technique in a novel informative way to analyze the foreign direct investment (FDI) inflow. This article examines the determinants of FDI inflow through IML with a supervised learning method to analyze the foreign investment determinants in Hungary by using an open-source artificial intelligence H2O platform. This author used three ML algorithms—general linear model (GML), gradient boosting machine (GBM), and random forest (RF) classifier—to analyze the FDI inflow from 2001 to 2018. The result of this study shows that in all three classifiers GBM performs better to analyze FDI inflow determinants. The variable value of production in a region is the most influenced determinant to the inflow of FDI in Hungarian regions. Explanatory visualizations are presented from the analyzed dataset, which leads to their use in decision-making.

A comprehensive understanding of the long-term data on the crop, soils, environment, climate, and production management would facilitate efficient data-driven decision-making in agriculture production under changing climate. We have employed an explainable machine learning algorithm (random forest model coupled with LIME; Local Interpretable Model-Agnostic Explanations framework) using multi-decadal (1981–2015) data on climate variables, soil properties, and yield of major crops across the Coterminous United States (CONUS). This data-driven approach explained the multi-faceted factors of crop production for corn, soybean, cotton, and wheat under field conditions by leveraging agricultural informatics. We attempted to show how crop yields can better be correlated and explained when production input varies along with changing climatic/environmental and edaphic conditions. Our findings suggest Growing Degree Days (GDDs) as important climatic factors, while water holding capacity is one of the dominant soil properties in interpreting crop yield variability. Our findings will facilitate growers, crop production scientists, land management specialists, stakeholders, and policy makers in their future decision-making processes related to sustainable and long-term soil, water, and crop management practices.

Objectives/Hypothesis: Obstructive sleep apnea (OSA) is associated with higher risk of morbidity and mortality related to cardiovascular disease (CVD). Due to overlapping clinical risk factors, identifying high‐risk patients with OSA who are likely to develop CVD remains challenging. We aimed to identify baseline clinical factors associated with the future development of CVD in patients with OSA. Study Design: Retrospective analysis of prospectively collected data. Methods: We performed a retrospective analysis of 967 adults aged 45 to 84 years and enrolled in the Multi‐Ethnic Study of Atherosclerosis. Six machine learning models were created using baseline clinical factors initially identified by stepwise variable selection. The performance of these models for the prediction of additional risk of CVD in OSA was calculated. Additionally, these models were evaluated for interpretability using locally interpretable model‐agnostic explanations. Results: Of the 967 adults without baseline OSA or CVD, 116 were diagnosed with OSA and CVD and 851 with OSA alone 10 years after enrollment. The best performing models included random forest (sensitivity 84%, specificity 99%, balanced accuracy 91%) and bootstrap aggregation (sensitivity 84%, specificity 100%, balanced accuracy 92%). The strongest predictors of OSA and CVD versus OSA alone were fasting glucose >91 mg/dL, diastolic pressure >73 mm Hg, and age >59 years. Conclusion: In the selected study population of adults without OSA or CVD at baseline, the strongest predictors of CVD in patients with OSA include fasting glucose, diastolic pressure, and age. These results may shape a strategy for cardiovascular risk stratification in patients with OSA and early intervention to mitigate CVD‐related morbidity. Level of Evidence: 3 Laryngoscope, 132:234–241, 2022 [ABSTRACT FROM AUTHOR]

In recent years, artificial intelligence-based computer aided diagnosis (CAD) system for the hepatitis has made great progress. Especially, the complex models such as deep learning achieve better performance than the simple ones due to the nonlinear hypotheses of the real world clinical data. However,complex model as a black box, which ignores why it make a certain decision, causes the model distrust from clinicians. To solve these issues, an explainable artificial intelligence (XAI) framework is proposed in this paper to give the global and local interpretation of auxiliary diagnosis of hepatitis while retaining the good prediction performance. First, a public hepatitis classification benchmark from UCI is used to test the feasibility of the framework. Then, the transparent and black-box machine learning models are both employed to forecast the hepatitis deterioration. The transparent models such as logistic regression (LR), decision tree (DT)and k-nearest neighbor (KNN) are picked. While the black-box model such as the eXtreme Gradient Boosting (XGBoost), support vector machine (SVM), random forests (RF) are selected. Finally, the SHapley Additive exPlanations (SHAP), Local Interpretable Model-agnostic Explanations (LIME) and Partial Dependence Plots (PDP) are utilized to improve the model interpretation of liver disease. The experimental results show that the complex models outperform the simple ones. The developed RF achieves the highest accuracy (91.9%) among all the models. The proposed framework combining the global and local interpretable methods improves the transparency of complex models, and gets insight into the judgments from the complex models, thereby guiding the treatment strategy and improving the prognosis of hepatitis patients. In addition, the proposed framework could also assist the clinical data scientists to design a more appropriate structure of CAD. [ABSTRACT FROM AUTHOR]

Cushing's disease (CD) is a rare disease that occurs in 1.2–1.4 persons per million population per year. Recurrence prediction after transsphenoidal surgery (TSS) is important for determining individual treatment and follow-up strategies. Between 2000 and 2017, 354 CD patients with initial postoperative remission and long-term follow-up data were enrolled from Peking union medical college hospital (PUMCH) to predict recurrence, and PUMCH is one of the largest CD treatment centers in the world. We first investigated the effect of a factorization machine (FM)-based neural network to predict recurrence after TSS for CD. This method could automatically reduce a portion of the cross-feature selection work with acceptable parameters. We conducted a performance comparison of various algorithms on the collected dataset. To address the lack of interpretability of neural network models, we also used the local interpretable model-agnostic explanations approach, which provides an explanation in the form of relevant features of the predicted results by approximating the model behavior of the variables in a local manner. Compared with existing methods, the DeepFM model obtained the highest AUC value (0.869) and the lowest log loss value (0.256). According to the importance of each feature, three top features for the DeepFM model were postoperative morning adrenocorticotropic hormone level, age, and postoperative morning serum cortisol nadir. In the post hoc explanation phase, the above-mentioned importance-leading features made a great contribution to the prediction probability. The results showed that deep learning-based models could better aid neurosurgeons in recurrence prediction after TTS for patients with CD, and could contribute to determining individual treatment strategies. [ABSTRACT FROM AUTHOR]

In advancement of interpretable machine learning (IML), this research proposes local interpretable model-agnostic explanations (LIME) as a new visualization technique in a novel informative way to analyze the foreign direct investment (FDI) infl ow. This article examines the determinants of FDI infl ow through IML with a supervised learning method to analyze the foreign investment determinants in Hungary by using an open-source artifi cial intelligence H2O platform. This author used three ML algorithms--general linear model (GML), gradient boosting machine (GBM), and random forest (RF) classifi er--to analyze the FDI infl ow from 2001 to 2018. The result of this study shows that in all three classifi ers GBM performs better to analyze FDI infl ow determinants. The variable value of production in a region is the most infl uenced determinant to the infl ow of FDI in Hungarian regions. Explanatory visualizations are presented from the analyzed dataset, which leads to their use in decision-making. [ABSTRACT FROM AUTHOR]

Breast Cancer (BC) is believed to be the cancer that occurs most frequently in women worldwide, taking the lives of it's the victims. In early diagnosis aids the patients to survive under greater probability. Several existing studies utilize diagnostic mechanisms via histopathology image for early identification of breast tumors. However, it increases the medical costs and consumes the time. Thus, in order to accurately classify the breast tumor, this study suggests a novel explainable DL technique. Using this technique, better accuracy is achieved while performing classifications. Improved accuracy may greatly help the medical practitioners for classifying breast cancer effectively. Initially, adaptive unsharp mask filtering (AUMF) technique is proposed to remove the noise and enhance the quality of the image. Finally, Explainable Soft Attentive EfficientNet (ESAE-Net) technique is introduced to classify the breast tumor (BT). Four explainable algorithms are investigated for improved visualizations over the BTs: Gradient-Weighted Class Activation Mapping (Grad-CAM) Shapley additive explanations (SHAP), Contextual Importance and Utility (CIU), and Local Interpretable Model-Agnostic Explanations (LIME). The suggested approach uses two publicly accessible images of breast histopathology and is carried out on a Python platform. Performance metrics such as time complexity, False Discovery Rate (FDR), accuracy, and Mathew's correlation coefficient (MCC) are examined and contrasted with traditional research. In the experimental section, the proposed obtains an accuracy of 97.85% for dataset 1 and accuracy of 98.05% for dataset 2. In comparison with other existing methods, the proposed method is more efficient while using ESAE-Net for classifying the Breast cancer. • A novel and effective XAI based breast cancer (BC) classification framework (ESAE-Net) that can provide better learning interpretations and decision making processes effectively. • To introduce a novel adaptive unsharp mask filtering (AUMF) technique for enhancing the quality of the breast histopathology image. • To propose a new and efficient Explainable Soft Attentive EfficientNet (ESAE-Net) model to deal with the classification of BC from histopathology images. • To improve the proposed ESAE-Net by introducing various interpretability procedures like LIME, SHAP, CIU and Grad-CAM. • To validate the proposed method by carrying out extensive simulations under different simulation scenarios to prove the effectiveness of the proposed framework compared to other XAI frameworks. [ABSTRACT FROM AUTHOR]

There is a lot of research on the neural models used for short-term load forecasting (STLF), which is crucial for improving the sustainable operation of energy systems with increasing technical, economic, and environmental requirements. Neural networks are computationally powerful; however, the lack of clear, readable and trustworthy justification of STLF obtained using such models is a serious problem that needs to be tackled. The article proposes an approach based on the local interpretable model-agnostic explanations (LIME) method that supports reliable premises justifying and explaining the forecasts. The use of the proposed approach makes it possible to improve the reliability of heuristic and experimental neural modeling processes, the results of which are difficult to interpret. Explaining the forecasting may facilitate the justification of the selection and the improvement of neural models for STLF, while contributing to a better understanding of the obtained results and broadening the knowledge and experience supporting the enhancement of energy systems security based on reliable forecasts and simplifying dispatch decisions.

Prediction with higher accuracy is vital for stock market prediction. Recently, considerable amount of effort has been poured into employing machine learning (ML) techniques for successfully predicting stock market price direction. No matter how successful the proposed prediction model is, it can be argued that there occur two major drawbacks for further increasing the prediction accuracy. The first one can be referred as the black box nature of ML techniques, in other words inference from the predictions cannot be explained. Furthermore, due to the complex characteristics of the predicted time series, no matter how sophisticated techniques are employed, it would be very difficult to achieve a marginal increase in accuracy that would meaningfully offset the additional computational burden it brings in. For these two reasons, instead of chasing incremental improvements in accuracy, we propose utilizing an "e X plainable A rtificial I ntelligence" (XAI) approach which can be employed for assessing the reliability of the predictions hence allowing decision maker to abstain from poor decisions which are responsible for declining overall prediction performance. If there would be a measure of how sure the prediction model is on any prediction, the predictions with a relatively higher reliability could be used to make a decision while lower quality decisions could be avoided. In this study, a novel two-stage stacking ensemble model for stock market direction prediction based on ML, empirical mode decomposition (EMD) and XAI is proposed. Our experiments have shown that, proposed prediction model supported with local interpretable model-agnostic explanations (LIME) achieved the highest accuracy of 0.9913 when only the most trusted predictions have been considered on KOSPI dataset and analogous successful results have been obtained from five other major stock market indices. • Accuracy increase in stock market prediction model when cannot be increased further with explainable machine learning. • Novel application method for model explainability with machine learning prediction models. • Performance increase for Empirical Mode Decomposition based prediction models with ensemble learning. [ABSTRACT FROM AUTHOR]

• Two thousand office building simulation models were generated using Latin hypercube sampling for various climate zones. • The XAI techniques, i.e., SHAP and LIME could identify the essential input variables more effectively compared with SRC. • The SHAP method appears to be superior over LIME method in identifying the most essential input variables. • The types and essential number of input parameters for the deep learning-based model varied with different climate zones. Simplification of input variables can increase the applicability of Artificial Intelligence (AI) in building load prediction. The most essential inputs for AI therefore need to be identified via a significance level test. In this study, the significance of the input parameters was evaluated using the standardized regression coefficient (SRC) and Explainable AI methods, i.e., Local Interpretable Model-agnostic Explanations (LIME) and SHapley Additive exPlanations (SHAP). To consider various types of office buildings operating under different climates, the constraints of the U.S. Department of Energy reference buildings with the Latin hypercube sampling were used to generate two thousand building models for each of five climate zones classified by the Köppen and Geiger system. The number and types of essential inputs for deep learning-based building load prediction were identified based on SRC, LIME, and SHAP. By comparison, the SHAP method gave the smallest number of essential inputs for accurate load prediction. In addition, the types of essential inputs varied with different climates. This study presented the potential elimination of weather sensors with time variables depending on the climate conditions. It would help building practitioners and non-experts to determine the essential inputs to build up a simple but accurate deep learning-based building system. [ABSTRACT FROM AUTHOR]

Explainable Artificial Intelligence (XAI) makes AI understandable to the human user particularly when the model is complex and opaque. Local Interpretable Model-agnostic Explanations (LIME) has an image explainer package that is used to explain deep learning models. The image explainer of LIME needs some parameters to be manually tuned by the expert in advance, including the number of top features to be seen and the number of superpixels in the segmented input image. This parameter tuning is a time-consuming task. Hence, with the aim of developing an image explainer that automizes image segmentation, this paper proposes Ensemble-based Genetic Algorithm Explainer (EGAE) for melanoma cancer detection that automatically detects and presents the informative sections of the image to the user. EGAE has three phases. First, the sparsity of chromosomes in GAs is determined heuristically. Then, multiple GAs are executed consecutively. However, the difference between these GAs are in different number of superpixels in the input image that result in different chromosome lengths. Finally, the results of GAs are ensembled using consensus and majority votings. This paper also introduces how Euclidean distance can be used to calculate the distance between the actual explanation (delineated by experts) and the calculated explanation (computed by the explainer) for accuracy measurement. Experimental results on a melanoma dataset show that EGAE automatically detects informative lesions, and it also improves the accuracy of explanation in comparison with LIME efficiently. The python codes for EGAE, the ground truths delineated by clinicians, and the melanoma detection dataset are available at https://github.com/KhaosResearch/EGAE., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2023 The Author(s). Published by Elsevier Ltd.. All rights reserved.)

There is a lot of research on the neural models used for short-term load forecasting (STLF), which is crucial for improving the sustainable operation of energy systems with increasing technical, economic, and environmental requirements. Neural networks are computationally powerful; however, the lack of clear, readable and trustworthy justification of STLF obtained using such models is a serious problem that needs to be tackled. The article proposes an approach based on the local interpretable model-agnostic explanations (LIME) method that supports reliable premises justifying and explaining the forecasts. The use of the proposed approach makes it possible to improve the reliability of heuristic and experimental neural modeling processes, the results of which are difficult to interpret. Explaining the forecasting may facilitate the justification of the selection and the improvement of neural models for STLF, while contributing to a better understanding of the obtained results and broadening the knowledge and experience supporting the enhancement of energy systems security based on reliable forecasts and simplifying dispatch decisions. [ABSTRACT FROM AUTHOR]

Developing high-efficient and green energy-saving stirred tank reactors is of vital importance in liquid-liquid processing industries. Therefore, this work revealed the energy consumption, flow characteristics, and explored the energy-saving strategies of novel-designed Cup-shape Blade (CB) stirred tank reactors characterized by easy fabrication, low energy consumption and large pumping flow rate. Computational Fluid Dynamics (CFD) and Artificial Neural Network (ANN) were conducted to investigate the influences of geometrical parameters on liquid-liquid power number (N P), flow number (N Q) and flow patterns. The improved method for hyper-parameter specification indicates that the Bayesian regulation training algorithm and the two-hidden-layer scheme possess the best performance. Back Propagation Neural Network (BPNN) presents the best predictive accuracy for N P and N Q , and Learning Vector Quantization (LVQ) network shows the best classification capability of flow patterns. Analyses based on the ANN modeling and the "Local Interpretable Model-agnostic Explanations" method show the joint and complicated effects of geometrical parameters on energy consumption and flow characteristics. Procedures were proposed to ascertain the energy-efficiency blade configuration based on ANN. [Display omitted] • CFD and ANN method were used to study the performance of CB stirred tanks. • An optimized method was adopted to ascertain the details of ANN. • Local sensitivities of input variables were implemented. • The procedures to select an energy-efficient blade configuration were proposed. [ABSTRACT FROM AUTHOR]