Search

Background: Shapley values have been used extensively in machine learning, not only to explain black box machine learning models, but among other tasks, also to conduct model debugging, sensitivity and fairness analyses and to select important features for robust modelling and for further follow-up analyses. Shapley values satisfy certain axioms that promote fairness in distributing contributions of features toward prediction or reducing error, after accounting for non-linear relationships and interactions when complex machine learning models are employed. Recently, feature selection methods using predictive Shapley values and p-values have been introduced, including powershap. Methods: We present a novel feature selection method, LLpowershap, that takes forward these recent advances by employing loss-based Shapley values to identify informative features with minimal noise among the selected sets of features. We also enhance the calculation of p-values and power to identify informative features and to estimate number of iterations of model development and testing. Results: Our simulation results show that LLpowershap not only identifies higher number of informative features but outputs fewer noise features compared to other state-of-the-art feature selection methods. Benchmarking results on four real-world datasets demonstrate higher or comparable predictive performance of LLpowershap compared to other Shapley based wrapper methods, or filter methods. LLpowershap is also ranked the best in mean ranking among the seven feature selection methods tested on the benchmark datasets. Conclusion: Our results demonstrate that LLpowershap is a viable wrapper feature selection method that can be used for feature selection in large biomedical datasets and other settings. [ABSTRACT FROM AUTHOR]

Achieving sustainable agricultural advancements necessitates optimizing crop yields while maintaining environmental stewardship. Our research addresses this critical imperative by introducing an innovative predictive model that refines crop recommendation systems through advanced machine learning techniques, specifically random forest and SHapley Additive exPlanations (SHAP). This study aims to overcome the limitations of traditional advisory approaches by incorporating interpretability tools, clarifying the model's decision‐making process around specific instances. To enhance the model's local interpretability, we incorporated local interpretable model‐agnostic explanations (LIMEs), providing transparent explanations for each crop recommendation, which fosters user trust, particularly when predictions diverge from established expert opinions. We conducted our empirical investigation using a comprehensive dataset that includes various agricultural parameters, historical crop yields, and environmental conditions to evaluate the model's performance. Our findings indicate a significant improvement in predictive accuracy over traditional methods. The application of SHAP values offers a groundbreaking analysis of feature importance, enabling a precise quantification of the contributions of factors such as soil quality, climatic variables, and historical crop performance to the predictive outcomes. This research advances the field of precision agriculture by presenting a model that excels not only in accuracy but also in providing actionable insights through enhanced interpretability. By balancing advanced predictive capabilities with user‐centric explanations, our model represents a substantial step forward in developing data‐driven, transparent, and trustworthy agricultural advisories. [ABSTRACT FROM AUTHOR]

We explore the potential of machine learning (ML) models applied in two financial risk management areas, i.e., credit risk management and financial risk hedging, through two practical use cases. This comparative study starts with the issue of explainability in complex ML models used in peer-to-peer lending for credit risk management. The first use case examines the limitations of using Kernel-SHAP with dependent features and evaluates different methods for estimating these dependencies using the Lending Club dataset. Our results suggest that accounting for feature dependence improves the understanding and robustness of prediction explanations. The second use case investigates a dynamic method for hedging foreign exchange risk in international equity portfolios, emphasizing the importance of accurate forecasts of currency returns. The analysis demonstrates that predictions yielded by ML models can significantly enhance the hedging of portfolios against currency risk. These findings highlight the transformative potential of advanced ML models in financial risk management, showcasing their capability to improve financial risk measurement and management. Further, our study outlines future research directions to advance this field. [ABSTRACT FROM AUTHOR]

Complex Machine Learning (ML) models used to support decision-making in peer-to-peer (P2P) lending often lack clear, accurate, and interpretable explanations. While the game-theoretic concept of Shapley values and its computationally efficient variant Kernel SHAP may be employed for this aim, similarly to other existing methods, the latter makes the assumption that the features are independent. The assumption of uncorrelated features in credit risk management is fairly restrictive and, thus, prediction explanations coming from correlated features might result in highly misleading Shapley values, even when considering simple models. We therefore propose an evaluation of different dependent-feature estimation methods of Kernel SHAP for classification purposes in credit risk management. We show that dependent-feature estimation of Shapley values can improve the understanding of true prediction explanations, their robustness and is essential for better identifying the most relevant variables to default predictions coming from black-box ML models. RESEARCH HIGHLIGHTS: We propose estimation of feature-dependent Shapley values for P2P credit risk management We consider different linear and non-linear predictive models with varying degrees of dependence Dependent feature estimation of Shapley values can improve prediction explanations and their robustness Loan amount and interest rate are the most determinant features to loan default prediction explanations [ABSTRACT FROM AUTHOR]

Besides its structural and economic characteristics, the location of a property is probably one of the most important determinants of its underlying value. In contrast to property valuations, there are hardly any approaches to date that evaluate the quality of a real estate location in an automated manner. The reasons are the complexity, the number of interactions and the non-linearities underlying the quality specifications of a certain location. By combining a state-of-the-art machine learning algorithm and the local post-hoc model agnostic method of Shapley Additive Explanations, this paper introduces a newly developed approach – called SHAP location score – that is able to detect these complexities and enables assessing real estate locations in a data-based manner. The SHAP location score represents an intuitive and flexible approach based on econometric modeling techniques and the basic assumptions of hedonic pricing theory. The approach can be applied post-hoc to any common machine learning method and can be flexibly adapted to the respective needs. This constitutes a significant extension of traditional urban models and offers many advantages for a wide range of real estate players. [ABSTRACT FROM AUTHOR]

Featured Application: We show illustrative examples of sexist language to describe the taxonomy and explainability analysis. In this study, we present a new approach that combines multiple Bidirectional Encoder Representations from Transformers (BERT) architectures with a Convolutional Neural Network (CNN) framework designed for sexism detection in text at a granular level. Our method relies on the analysis and identification of the most important terms contributing to sexist content using Shapley Additive Explanations (SHAP) values. This approach involves defining a range of Sexism Scores based on both model predictions and explainability, moving beyond binary classification to provide a deeper understanding of the sexism-detection process. Additionally, it enables us to identify specific parts of a sentence and their respective contributions to this range, which can be valuable for decision makers and future research. In conclusion, this study introduces an innovative method for enhancing the clarity of large language models (LLMs), which is particularly relevant in sensitive domains such as sexism detection. The incorporation of explainability into the model represents a significant advancement in this field. The objective of our study is to bridge the gap between advanced technology and human comprehension by providing a framework for creating AI models that are both efficient and transparent. This approach could serve as a pipeline for future studies to incorporate explainability into language models. [ABSTRACT FROM AUTHOR]

This work proposes a method to rank the contribution of containment measures, vaccination coverage and mobility to contain the evolution of the COVID-19 pandemic in different states of Brazil. The proposed method applies the automatic learning of regression models using decision trees through the XGBoost algorithm. To interpret it globally, the SHapley Additive exPlanations (SHAP) was used, which is an algorithm based on Shapley's cooperative game theory to quantify the contribution of the analyzed characteristics to the evolution of the target variable (deaths). The evaluation results of the method indicated its efficiency to quantify the contribution of each variable in a robust way. It reveals that the percentages of vaccine coverage of the first and second dose, in addition to the closure of schools, were the measures that had the greatest contribution to the evolution of the number of deaths from COVID-19. The weighting of the variables can help the responsible actors in the elaboration of public policies to minimize the socioeconomic effects in their regions. Since, in countries with great social inequality, the use of only a few more efficient measures would be less harmful than a lockdown, as it would be extremely harmful to the quality of life of these poorer populations.

The synthesis of silver nanoparticles with controlled physicochemical properties is essential for governing their intended functionalities and safety profiles. However, synthesis process involves multiple parameters that could influence the resulting properties. This challenge could be addressed with the development of predictive models that forecast endpoints based on key synthesis parameters. In this study, we manually extracted synthesis-related data from the literature and leveraged various machine learning algorithms. Data extraction included parameters such as reactant concentrations, experimental conditions, as well as physicochemical properties. The antibacterial efficiencies and toxicological profiles of the synthesized nanoparticles were also extracted. In a second step, based on data completeness, we employed regression algorithms to establish relationships between synthesis parameters and desired endpoints and to build predictive models. The models for core size and antibacterial efficiency were trained and validated using a cross-validation approach. Finally, the features’ impact was evaluated via Shapley values to provide insights into the contribution of features to the predictions. Factors such as synthesis duration, scale of synthesis and the choice of capping agents emerged as the most significant predictors. This study demonstrated the potential of machine learning to aid in the rational design of synthesis process and paves the way for the safe-by-design principles development by providing insights into the optimization of the synthesis process to achieve the desired properties. Finally, this study provides a valuable dataset compiled from literature sources with significant time and effort from multiple researchers. Access to such datasets notably aids computational advances in the field of nanotechnology.

Abstract Background Shapley values have been used extensively in machine learning, not only to explain black box machine learning models, but among other tasks, also to conduct model debugging, sensitivity and fairness analyses and to select important features for robust modelling and for further follow-up analyses. Shapley values satisfy certain axioms that promote fairness in distributing contributions of features toward prediction or reducing error, after accounting for non-linear relationships and interactions when complex machine learning models are employed. Recently, feature selection methods using predictive Shapley values and p-values have been introduced, including powershap. Methods We present a novel feature selection method, LLpowershap, that takes forward these recent advances by employing loss-based Shapley values to identify informative features with minimal noise among the selected sets of features. We also enhance the calculation of p-values and power to identify informative features and to estimate number of iterations of model development and testing. Results Our simulation results show that LLpowershap not only identifies higher number of informative features but outputs fewer noise features compared to other state-of-the-art feature selection methods. Benchmarking results on four real-world datasets demonstrate higher or comparable predictive performance of LLpowershap compared to other Shapley based wrapper methods, or filter methods. LLpowershap is also ranked the best in mean ranking among the seven feature selection methods tested on the benchmark datasets. Conclusion Our results demonstrate that LLpowershap is a viable wrapper feature selection method that can be used for feature selection in large biomedical datasets and other settings.

Electricity expense management presents significant challenges, as this resource is susceptible to various influencing factors. In universities, the demand for this resource is rapidly growing with institutional expansion and has a significant environmental impact. In this study, the machine learning models long short-term memory (LSTM), random forest (RF), support vector regression (SVR), and extreme gradient boosting (XGBoost) were trained with historical consumption data from the Federal Institute of Paraná (IFPR) over the last seven years and climatic variables to forecast electricity consumption 12 months ahead. Datasets from two campuses were adopted. To improve model performance, feature selection was performed using Shapley additive explanations (SHAP), and hyperparameter optimization was carried out using genetic algorithm (GA) and particle swarm optimization (PSO). The results indicate that the proposed cooperative ensemble learning approach named Weaker Separator Booster (WSB) exhibited the best performance for datasets. Specifically, it achieved an sMAPE of 13.90% and MAE of 1990.87 kWh for the IFPR–Palmas Campus and an sMAPE of 18.72% and MAE of 465.02 kWh for the Coronel Vivida Campus. The SHAP analysis revealed distinct feature importance patterns across the two IFPR campuses. A commonality that emerged was the strong influence of lagged time-series values and a minimal influence of climatic variables.

In the field of geomatics, artificial intelligence (AI) and especially machine learning (ML) are rapidly transforming the field of geomatics with respect to collecting, managing, and analyzing spatial data. Feature selection as a building block in ML is crucial because it directly impacts the performance and predictive power of a model by selecting the most critical variables and eliminating the redundant and irrelevant ones. Random forests have now been used for decades and allow for building models with high accuracy. However, finding the most expressive features from the dataset by selecting the most important features within random forests is still a challenging question. The often-used internal Gini importances of random forests are based on the amount of training examples that are divided by a feature but fail to acknowledge the magnitude of change in the target variable, leading to suboptimal selections. Shapley values are an established and unified framework for feature attribution, i.e., specifying how much each feature in a trained ML model contributes to the predictions for a given instance. Previous studies highlight the effectiveness of Shapley values for feature selection in real-world applications, while other research emphasizes certain theoretical limitations. This study provides an application-driven discussion of Shapley values for feature selection by first proposing four necessary conditions for a successful feature selection with Shapley values that are extracted from a multitude of critical research in the field. Given these valuable conditions, Shapley value feature selection is nevertheless a model averaging procedure by definition, where unimportant features can alter the final selection. Therefore, we additionally present Conditional Feature Selection (CFS) as a novel algorithm for performing feature selection that mitigates this problem and use it to evaluate the impact of model averaging in several real-world examples, covering the use of ML in geomatics. The results of this study show Shapley values as a good measure for feature selection when compared with Gini feature importances on four real-world examples, improving the RMSE by 5% when averaged over selections of all possible subset sizes. An even better selection can be achieved by CFS, improving on the Gini selection by approximately 7.5% in terms of RMSE. For random forests, Shapley value calculation can be performed in polynomial time, offering an advantage over the exponential runtime of CFS, building a trade-off to the lost accuracy in feature selection due to model averaging.

Electricity expense management presents significant challenges, as this resource is susceptible to various influencing factors. In universities, the demand for this resource is rapidly growing with institutional expansion and has a significant environmental impact. In this study, the machine learning models long short-term memory (LSTM), random forest (RF), support vector regression (SVR), and extreme gradient boosting (XGBoost) were trained with historical consumption data from the Federal Institute of Paraná (IFPR) over the last seven years and climatic variables to forecast electricity consumption 12 months ahead. Datasets from two campuses were adopted. To improve model performance, feature selection was performed using Shapley additive explanations (SHAP), and hyperparameter optimization was carried out using genetic algorithm (GA) and particle swarm optimization (PSO). The results indicate that the proposed cooperative ensemble learning approach named Weaker Separator Booster (WSB) exhibited the best performance for datasets. Specifically, it achieved an sMAPE of 13.90% and MAE of 1990.87 kWh for the IFPR–Palmas Campus and an sMAPE of 18.72% and MAE of 465.02 kWh for the Coronel Vivida Campus. The SHAP analysis revealed distinct feature importance patterns across the two IFPR campuses. A commonality that emerged was the strong influence of lagged time-series values and a minimal influence of climatic variables. [ABSTRACT FROM AUTHOR]

Detecting and reacting to non-technical losses (NTL) is a fundamental activity that energy providers need to face in their daily routines. This is known to be challenging since the phenomenon of NTL is multi-factored, dynamic and extremely contextual, which makes artificial intelligence (AI) and, in particular, machine learning, natural areas to bring effective and tailored solutions. If the human factor is disregarded in the process of detecting NTL, there is a high risk of performance degradation since typical problems like dataset shift and biases cannot be easily identified by an algorithm. This paper presents a case study on incorporating explainable AI (XAI) in a mature NTL detection system that has been in production in the last years both in electricity and gas. The experience shows that incorporating this capability brings interesting improvements to the initial system and especially serves as a common ground where domain experts, data scientists, and business analysts can meet. [ABSTRACT FROM AUTHOR]

This paper proposes a novel approach to explain the predictions made by data-driven methods. Since such predictions rely heavily on the data used for training, explanations that convey information about how the training data affects the predictions are useful. The paper proposes a novel approach to quantify how different data-clusters of the training data affect a prediction. The quantification is based on Shapley values, a concept which originates from coalitional game theory, developed to fairly distribute the payout among a set of cooperating players. A player's Shapley value is a measure of that player's contribution. Shapley values are often used to quantify feature importance, ie. how features affect a prediction. This paper extends this to cluster importance, letting clusters of the training data act as players in a game where the predictions are the payouts. The novel methodology proposed in this paper lets us explore and investigate how different clusters of the training data affect the predictions made by any black-box model, allowing new aspects of the reasoning and inner workings of a prediction model to be conveyed to the users. The methodology is fundamentally different from existing explanation methods, providing insight which would not be available otherwise, and should complement existing explanation methods, including explanations based on feature importance. [ABSTRACT FROM AUTHOR]

In the field of geomatics, artificial intelligence (AI) and especially machine learning (ML) are rapidly transforming the field of geomatics with respect to collecting, managing, and analyzing spatial data. Feature selection as a building block in ML is crucial because it directly impacts the performance and predictive power of a model by selecting the most critical variables and eliminating the redundant and irrelevant ones. Random forests have now been used for decades and allow for building models with high accuracy. However, finding the most expressive features from the dataset by selecting the most important features within random forests is still a challenging question. The often-used internal Gini importances of random forests are based on the amount of training examples that are divided by a feature but fail to acknowledge the magnitude of change in the target variable, leading to suboptimal selections. Shapley values are an established and unified framework for feature attribution, i.e., specifying how much each feature in a trained ML model contributes to the predictions for a given instance. Previous studies highlight the effectiveness of Shapley values for feature selection in real-world applications, while other research emphasizes certain theoretical limitations. This study provides an application-driven discussion of Shapley values for feature selection by first proposing four necessary conditions for a successful feature selection with Shapley values that are extracted from a multitude of critical research in the field. Given these valuable conditions, Shapley value feature selection is nevertheless a model averaging procedure by definition, where unimportant features can alter the final selection. Therefore, we additionally present Conditional Feature Selection (CFS) as a novel algorithm for performing feature selection that mitigates this problem and use it to evaluate the impact of model averaging in several real-world examples, covering the use of ML in geomatics. The results of this study show Shapley values as a good measure for feature selection when compared with Gini feature importances on four real-world examples, improving the RMSE by 5% when averaged over selections of all possible subset sizes. An even better selection can be achieved by CFS, improving on the Gini selection by approximately 7.5% in terms of RMSE. For random forests, Shapley value calculation can be performed in polynomial time, offering an advantage over the exponential runtime of CFS, building a trade-off to the lost accuracy in feature selection due to model averaging. [ABSTRACT FROM AUTHOR]

Feature selection is one of the most relevant processes in any methodology for creating a statistical learning model. Usually, existing algorithms establish some criterion to select the most influential variables, discarding those that do not contribute to the model with any relevant information. This methodology makes sense in a static situation where the joint distribution of the data does not vary over time. However, when dealing with real data, it is common to encounter the problem of the dataset shift and, specifically, changes in the relationships between variables (concept shift). In this case, the influence of a variable cannot be the only indicator of its quality as a regressor of the model, since the relationship learned in the training phase may not correspond to the current situation. In tackling this problem, our approach establishes a direct relationship between the Shapley values and prediction errors, operating at a more local level to effectively detect the individual biases introduced by each variable. The proposed methodology is evaluated through various examples, including synthetic scenarios mimicking sudden and incremental shift situations, as well as two real-world cases characterized by concept shifts. Additionally, we perform three analyses of standard situations to assess the algorithm's robustness in the absence of shifts. The results demonstrate that our proposed algorithm significantly outperforms state-of-the-art feature selection methods in concept shift scenarios, while matching the performance of existing methodologies in static situations. [ABSTRACT FROM AUTHOR]

Background: Haemoperfusion (HP) is an innovative extracorporeal therapy that utilizes special cartridges to filter the blood, effectively removing pro-inflammatory cytokines, toxins, and pathogens in COVID-19 patients. This retrospective cohort study aimed to assess the clinical benefits of HP for severe COVID-19 cases using Shapley values for machine learning models. Methods: The research involved 578 inpatients (≥ 20 years old) admitted to Baqiyatallah hospital (Tehran, Iran). The control group (359 patients) received standard treatment, including high doses of corticosteroids (a single 500 mg methylprednisolone pulse, followed by 250 mg for 2 days), categorized as regimen (I). On the other hand, the HP group (219 patients) received regimen II, consisting of the same corticosteroid treatment (regimen I) along with haemoperfusion using Cytosorb H300. The frequency of haemoperfusion sessions varied based on the type of lung involvement determined by chest CT scans. In addition, the value function v defines the Shapley value of the i th feature for the query point x , where the input matrix features represent individual characteristics, drugs, and history and clinical conditions of the patient. Results: Our data showed a favorable clinical response in the HP group compared to the control group. Notably, one-to-three sessions of HP using the CytoSorb® 300 cartridge led to reduced ventilation requirements and mortality rates in severe COVID-19 patients. Shapley values were calculated to evaluate the contribution of haemoperfusion among other factors, such as side effects, medications, and individual characteristics, to COVID-19 patient outcomes. In addition, there is a significant difference between the two groups among the treatments and medications used remdesivir, adalimumab, tocilizumab, favipiravir, Interferon beta-1a, enoxaparin prophylaxis, enoxaparin full dose, heparin prophylaxis, and heparin full dose (P < 0.05). It seems that haemoperfusion has a positive impact on the reduction of inflammation markers and renal functional such as ferritin and creatinine, respectively, as well as d-dimer and WBC levels in the HP group were significantly lower than the control group. Conclusion: The findings indicated that haemoperfusion played a crucial role in predicting patient survival, making it a significant feature in classifying patients' prognoses. [ABSTRACT FROM AUTHOR]

Social interactions are essential for well-being. Therefore, researchers increasingly attempt to capture an individual's social context to predict well-being, including mood. Different tools are used to measure various aspects of the social context. Digital phenotyping is a commonly used technology to assess a person's social behavior objectively. The experience sampling method (ESM) can capture the subjective perception of specific interactions. Lastly, egocentric networks are often used to measure specific relationship characteristics. These different methods capture different aspects of the social context over different time scales that are related to well-being, and combining them may be necessary to improve the prediction of well-being. Yet, they have rarely been combined in previous research. To address this gap, our study investigates the predictive accuracy of mood based on the social context. We collected intensive within-person data from multiple passive and self-report sources over a 28–day period in a student sample (Participants: N = 11, ESM measures: N = 1313). We trained individualized random forest machine learning models, using different predictors included in each model summarized over different time scales. Our findings revealed that even when combining social interactions data using different methods, predictive accuracy of mood remained low. The average coefficient of determination over all participants was 0.06 for positive and negative affect and ranged from − 0.08 to 0.3, indicating a large amount of variance across people. Furthermore, the optimal set of predictors varied across participants; however, predicting mood using all predictors generally yielded the best predictions. While combining different predictors improved predictive accuracy of mood for most participants, our study highlights the need for further work using larger and more diverse samples to enhance the clinical utility of these predictive modeling approaches. [ABSTRACT FROM AUTHOR]

Shapley values originated in cooperative game theory but are extensively used today as a model-agnostic explanation framework to explain predictions made by complex machine learning models in the industry and academia. There are several algorithmic approaches for computing different versions of Shapley value explanations. Here, we consider Shapley values incorporating feature dependencies, referred to as conditional Shapley values, for predictive models fitted to tabular data. Estimating precise conditional Shapley values is difficult as they require the estimation of non-trivial conditional expectations. In this article, we develop new methods, extend earlier proposed approaches, and systematize the new refined and existing methods into different method classes for comparison and evaluation. The method classes use either Monte Carlo integration or regression to model the conditional expectations. We conduct extensive simulation studies to evaluate how precisely the different method classes estimate the conditional expectations, and thereby the conditional Shapley values, for different setups. We also apply the methods to several real-world data experiments and provide recommendations for when to use the different method classes and approaches. Roughly speaking, we recommend using parametric methods when we can specify the data distribution almost correctly, as they generally produce the most accurate Shapley value explanations. When the distribution is unknown, both generative methods and regression models with a similar form as the underlying predictive model are good and stable options. Regression-based methods are often slow to train but quickly produce the Shapley value explanations once trained. The vice versa is true for Monte Carlo-based methods, making the different methods appropriate in different practical situations. [ABSTRACT FROM AUTHOR]

Hepatitis C virus (HCV) infection remains a significant global health challenge, often resulting in severe long-term physical complexity and even death. Since its discovery, HCV has exhibited substantial genetic variability, complicating vaccine development. Although some therapeutic approach have shown efficacy against certain HCV genotypes, a universally effective vaccine is still lacking. Recent research suggests that the body's cellular immune response, particularly T cell epitopes of HCV (TCE-HCVs), plays a vital role in fighting the virus. Therefore, the precise and rapid identification of TCE-HCVs is essential for chronic HCV infection. In this work, we proposed a novel TCE-HCVs prediction model AttLSTM, which combines attention mechanism and long short-term memory (LSTM). Specifically, we employed four robust feature encoding techniques: One-Hot Encoding, Global Vectors (GloVe), fastText, and Word2Vec to encode protein sequences. Additionally, k-mer embedding was utilized to help the model identify significant subsequence fragments within the protein sequences. To optimize the model's performance, irrelevant features are eliminated using the SHapley Additive exPlanations (SHAP) approach. The resulting optimal feature subset was then fed into the AttLSTM model to identify TCE-HCVs. The attention mechanism in this model dynamically captures the pairwise correlations of each neighboring target pair within a sliding window, thereby enhancing the understanding of the local environment of target residues. Extensive experiments showed that AttLSTM outperformed conventional machine learning (ML) classifiers in predictive performance. Notably, in k-fold cross-validation, AttLSTM achieved superior performance compared to existing methods with accuracy of 80.77 %, MCC of 0.632, and AUC of 0.891. This exceptional performance indicates that AttLSTM has a strong predictive capability for identifying TCE-HCVs. We anticipate that AttLSTM will expedite the rapid identification of promising TCE-HCVs, aiding in the development of diagnostic and immunotherapeutic treatments for HCV in the future. In addition, we have developed a web server for real-time prediction using our proposed model, which is available at http://attlstmhcv.xyz/.

Background: An improved understanding of the reasoning behind model decisions can enhance the use of machine learning (ML) models in credit scoring. Although ML models are widely regarded as highly accurate, the use of these models in settings that require explanation of model decisions has been limited because of the lack of transparency. Especially in the banking sector, model risk frameworks frequently require a significant level of model interpretability. Aim: The aim of the article is to evaluate traditional model risk frameworks to determine their appropriateness when validating ML models in credit scoring and enhance the use of ML models in regulated environments by introducing a ML interpretability technique in model validation frameworks. Setting: The research considers model risk frameworks and regulatory guidelines from various international institutions. Method: The research is qualitative in nature and shows how through integrating traditional and non-traditional model risk frameworks, the practitioner can leverage trusted techniques and extend traditional frameworks to address key principles such as transparency. Results: The article proposes a model risk framework that utilises Shapley values to improve the explainability of ML models in credit scoring. Practical validation tests are proposed to enable transparency of model input variables in the validation process of ML models. Conclusion: Our results show that one can formulate a comprehensive validation process by integrating traditional and non-traditional frameworks. Contribution: This study contributes to existing model risk literature by proposing a new model validation framework that utilises Shapley values to explain ML model predictions in credit scoring.

Motor vehicle crashes are the leading cause of the death of teenagers in the United States. Young drivers have shown a higher propensity to get involved in crashes due to using a cellphone while driving, breaking the speed limit, and reckless driving. This study analyzed motor vehicle crashes involving young drivers using New Jersey crash data. Specifically, four years of crash data (2016–2019) were gathered and analyzed. Different machine learning (ML) methods, such as Random Forest, Light GBM, Catboost, and XGBoost, were used to predict the injury severity. The performance of the models was evaluated using accuracy, precision, and recall scores. In addition, interpretable ML techniques like sensitivity analysis and Shapley values were conducted to assess the most influential factors' impacts on young driver-related crashes. The results revealed that XGBoost performed better than Random Forest, CatBoost, and LightGBM models in crash severity prediction. Results from the sensitivity analysis showed that multi-vehicle crashes, angular crashes, crashes at intersections, and dark-not-lit conditions had increased crash severity. A partial dependence plot of SHAP values revealed that speeding in clear weather had a higher likelihood of injury crashes, and multi-vehicle crashes at the intersection had more injury crashes. We expect that the results obtained from this study will help policymakers and practitioners take appropriate countermeasures to improve the safety of young drivers in New Jersey.

In DCE-MRI, the degree of contrast uptake in normal fibroglandular tissue, i.e., background parenchymal enhancement (BPE), is a crucial biomarker linked to breast cancer risk and treatment outcome. In accordance with the Breast Imaging Reporting & Data System (BI-RADS), it should be visually classified into four classes. The susceptibility of such an assessment to inter-reader variability highlights the urgent need for a standardized classification algorithm. In this retrospective study, the first post-contrast subtraction images for 27 healthy female subjects were included. The BPE was classified slice-wise by two expert radiologists. The extraction of radiomic features from segmented BPE was followed by dataset splitting and dimensionality reduction. The latent representations were then utilized as inputs to a deep neural network classifying BPE into BI-RADS classes. The network's predictions were elucidated at the radiomic feature level with Shapley values. The deep neural network achieved a BPE classification accuracy of 84 ± 2% (p-value < 0.00001). Most of the misclassifications involved adjacent classes. Different radiomic features were decisive for the prediction of each BPE class underlying the complexity of the decision boundaries. A highly precise and explainable pipeline for BPE classification was achieved without user- or algorithm-dependent radiomic feature selection. [ABSTRACT FROM AUTHOR]

There are many times when visuals and text will often disrupt each other in a single space and, as a result, make both more difficult to understand. Many methods exist to counteract this, such as placing text and visuals in specific locations in the given space. We have utilized these certain "combinations" to a great extent. Still, when there are special cases where these methods are less effective, finding an optimal combination as a replacement is more difficult. To see other patterns and techniques to assist in locating salient elements, we can use game theory and graph theory, as well as a small amount of cognitive neuroscience, to determine what combinations allow words and visuals to complement each other rather than compete. The research done in this paper is exemplified by bento grids, a recent design trend popularized by Apple. [ABSTRACT FROM AUTHOR]

Speech emotion recognition (SER) has gained significant attention due to its several application fields, such as mental health, education, and human-computer interaction. However, the accuracy of SER systems is hindered by high-dimensional feature sets that may contain irrelevant and redundant information. To overcome this challenge, this study proposes an iterative feature boosting approach for SER that emphasizes feature relevance and explainability to enhance machine learning model performance. Our approach involves meticulous feature selection and analysis to build efficient SER systems. In addressing our main problem through model explainability, we employ a feature evaluation loop with Shapley values to iteratively refine feature sets. This process strikes a balance between model performance and transparency, which enables a comprehensive understanding of the model's predictions. The proposed approach offers several advantages, including the identification and removal of irrelevant and redundant features, leading to a more effective model. Additionally, it promotes explainability, facilitating comprehension of the model's predictions and the identification of crucial features for emotion determination. The effectiveness of the proposed method is validated on the SER benchmarks Toronto emotional speech set (TESS), Berlin Database of Emotional Speech (EMO-DB), Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS), and Surrey Audio-Visual Expressed Emotion (SAVEE) dataset, outperforming state-of-the-art methods. These results highlight the potential of the proposed technique in developing accurate and explainable SER systems. To the best of our knowledge, this is the first work to incorporate model explainability into an SER framework. The source code of this paper is publicly available via this https://github.com/alaaNfissi/Unveiling-Hidden-Factors-Explainable-AI-for-Feature-Boosting-in-Speech-Emotion-Recognition. [ABSTRACT FROM AUTHOR]

Retail companies are greatly interested in performing continuous monitoring of purchase traces of customers, to identify weak customers and take the necessary actions to improve customer satisfaction and ensure their revenues remain unaffected. In this paper, we formulate the customer churn prediction problem as a Predictive Process Monitoring (PPM) problem to be addressed under possible dynamic conditions of evolving retail data environments. To this aim, we propose TSUNAMI as a PPM approach to monitor the customer loyalty in the retail sector. It processes online the sale receipt stream produced by customers of a retail business company and learns a deep neural model to early detect possible purchase customer traces that will outcome in future churners. In addition, the proposed approach integrates a mechanism to detect concept drifts in customer purchase traces and adapts the deep neural model to concept drifts. Finally, to make decisions of customer purchase monitoring explainable to potential stakeholders, we analyse Shapley values of decisions, to explain which characteristics of the customer purchase traces are the most relevant for disentangling churners from non-churners and how these characteristics have possibly changed over time. Experiments with two benchmark retail data sets explore the effectiveness of the proposed approach. [ABSTRACT FROM AUTHOR]

Machine learning (ML) has shown its effectiveness in handling multi-geoinformation. Yet, the black-box nature of ML algorithms has restricted their widespread adoption in the domain of mineral prospectivity mapping (MPM). In this paper, methods for interpreting ML model predictions are introduced to aid ML-based MPM, with the goal of extracting richer insights from the ML modeling of an exploration geochemical dataset. The partial dependence plot (PDP) and accumulated local effect (ALE) plot, along with the SHAP value analysis, were utilized to demonstrate the application of random forest (RF) modeling within both regression and classification frameworks. Initially, the random forest regression (RFR) model established the relationship between the concentrations of Au and those of elements such as As, Sb, and Hg in the study area, and from this model, the most important geochemical elements and their quantitative relationships with Au were revealed by their contributions in the modeling through PDP and ALE analyses. Secondly, the RF classification modeling established the relationships of mineralization occurrences (i.e., known mineral deposits) with geochemical elements (i.e., Au, As, Sb, Hg, Cu, Pb, Zn, and Ag), as did RFR modeling. The most important geochemical elements for indicating regional Au mineralization and the trajectories of PDP and ALE reached a consensus that As and Sb contributed the most, both in the regression and classification modeling, with regard to Au mineralization. Finally, the SHAP values illustrated the behavior of the training samples (i.e., known mineral deposits) in RF modeling, and the resulting prospectivity map was evaluated using receiver operating characteristics. [ABSTRACT FROM AUTHOR]

Geochemical pattern recognition has long been of interest for geologists to reveal geochemical anomalies associated with mineralization. In regional-scale exploration, geochemical anomalies are derived conventionally from stream sediment samples and processed in the form of vectors, resulting in row-wise outliers. However, geochemical anomalies derived through various means of pattern recognition have shown their limits in depicting complex geochemical distributions. In this paper, we propose to utilize the Shapley value, linked to the Mahalanobis distance (MD), and cell-wise outlier detection to facilitate the recognition of anomalous geochemical indicator elements. First, by considering the compositional nature of geochemical data, multivariate outliers are detected based on the MD in isometric log-ratio coordinates. Secondly, to quantify the contributions of individual elements to the outlyingness of an outlier, Shapley values are used to express the MDs of data as outlyingness contributions of single elements. Finally, cell-wise outlier detection is introduced to examine and quantify the outlyingness of each cell in a geochemical data matrix. The outlying cells serve as criteria for further recognition of element associations. By analysing the Shapley values of individual elements and the outlying cells in a geochemical data matrix, more information contained in multivariate outliers can be recognized. Using this proposed methodology, the element associations that relate to regional mineralization in the study area were Au-only anomalies, Au–As–Sb anomalies, As–Sb–Hg anomalies and Ag-related anomalies. [ABSTRACT FROM AUTHOR]

Understanding the intricate relationships between climate and vegetation remains a fundamental challenge in contemporary ecology. The ability to anticipate the specific climatic factors affecting different tree species and understand how they respond is crucial for mitigating the impacts of climate change on forested ecosystems. Additionally, quantitatively assessing habitat loss resulting from anthropogenic activities is essential for informed conservation efforts. Our objective is to evaluate the potential distribution of pitch pine (Pinus rigida) in North America and assess the associated habitat loss. To achieve this, we employ a stepwise multidimensional climate envelope modeling approach, comparing two data-intensive models—the Variable Interaction Model (VIM) and the Variable Non-Interaction Model (VNM). These models discern the influence of diverse combinations of climatic characteristics on the distribution of the species. Both VNM and VIM employ Shapley values for factor ranking during construction. VNM assumes independent effects, resulting in a hyperrectangle-shaped climate envelope, while VIM considers interactions, yielding a complex, data-driven multidimensional envelope. Data integration involves mining the US Forest Inventories and climatic data, encompassing 19 parameters. The results unequivocally highlight the superior predictive accuracy of VIM compared to the Variable Non-Interaction Model, VNM. The modeling approach developed in this study has the potential to enhance species distribution models for various tree species in the context of evolving climatic conditions. [ABSTRACT FROM AUTHOR]

Fruit cracking is a preharvest physiological rind disorder in citrus, sometimes causing considerable yield loss. In recent years, reports from Israel and other countries suggest that cracking incidence has increased, which might indicate that climate change intensifies the phenomena. The study aims to develop a machine learning (ML) model for predicting the effect of climate measures (i.e., temperature, radiation, and humidity with daily resolution) along with management and environmental characteristics in two citrus mandarins, 'Nova' and 'Ori', one is prone to cracking and the other is less sensitive. ML model was developed based on data from approximately 250 citrus orchards across Israel collected over three seasons from 2019 to 2021. Our approach uses TSFRESH to extract and select features and SHAP (SHapley Additive exPlanations) to explain the factor's intensity using trained classification and regression models based on the H2O-AutoML package. Gathered data skewed toward a low cracking percentage better predicted low and medium cracking levels, with a classification accuracy of 76% and regression mean absolute error (MAE) of 4.78%. Our study reaffirms the genetic background's primary role in cracking. Notably, our analysis unveils fresh insights into cracking causes needing further exploration. The 40% quantile temperature (23.5 °C) is a novel finding as a learned threshold. 'Nova' may elevate cracking by 10%, 'Ori' could reduce it by 4%. Additionally, tree age exhibits a linear correlation when trees over 20 years correlate with up to 4% less cracking. These insights are crucial for comprehending, addressing, and managing the phenomenon at a significant spatial scale. The model, with further data support, may provide farmers with an effective tool for treating the severity of cracking incidence by developing a spatial–temporal decision-support system as a protocol to reduce the phenomenon on a regional scale and selecting regions that are relevant for citrus plantations. [ABSTRACT FROM AUTHOR]

Cooperative alternatives need complex multi-criteria decision-making (MCDM) consideration, especially in resource allocation, where the alternatives exhibit interdependent relationships. Traditional MCDM methods like the Analytic Hierarchy Process (AHP) and Analytic Network Process (ANP) often overlook the synergistic potential of cooperative alternatives. This study introduces a novel method integrating AHP/ANP with Shapley values, specifically designed to address this gap by evaluating alternatives on individual merits and their contributions within coalitions. Our methodology begins with defining problem structures and applying AHP/ANP to determine the criteria weights and alternatives' scores. Subsequently, we compute Shapley values based on coalition values, synthesizing these findings to inform resource allocation decisions more equitably. A numerical example of budget allocation illustrates the method's efficacy, revealing significant insights into resource distribution when cooperative dynamics are considered. Our results demonstrate the proposed method's superiority in capturing the nuanced interplay between criteria and alternatives, leading to more informed urban planning decisions. This approach marks a significant advancement in MCDM, offering a comprehensive framework that incorporates both the analytical rigor of AHP/ANP and the equitable considerations of cooperative game theory through Shapley values. [ABSTRACT FROM AUTHOR]

Introduction: The research in consumer neuroscience has identified computational methods, particularly artificial intelligence (AI) and machine learning, as a significant frontier for advancement. Previously, we utilized functional magnetic resonance imaging (fMRI) and artificial neural networks (ANNs) to model brain processes related to brand preferences in a paradigm exempted from motor actions. In the current study, we revisit this data, introducing recent advancements in explainable artificial intelligence (xAI) to gain insights into this domain. By integrating fMRI data analysis, machine learning, and xAI, our study aims to search for functional brain networks that support brand perception and, ultimately, search for brain networks that disentangle between preferred and indifferent brands, focusing on the early processing stages. Methods: We applied independent component analysis (ICA) to overcome the expected fMRI data's high dimensionality, which raises hurdles in AI applications. We extracted pertinent features from the returned ICs. An ANN is then trained on this data, followed by pruning and retraining processes. We then apply explanation techniques, based on path-weights and Shapley values, to make the network more transparent, explainable, and interpretable, and to obtain insights into the underlying brain processes. Results: The fully connected ANN model obtained an accuracy of 54.6%, which dropped to 50.4% after pruning. However, the retraining process allowed it to surpass the fully connected network, achieving an accuracy of 55.9%. The pathweights and Shapley-based analysis concludes that, regarding brand perception, the expected initial participation of the primary visual system is followed. Other brain areas participate in early processing and discriminate between preferred and indifferent brands, such as the cuneal and the lateral occipital cortices. Discussion: The most important finding is that a split between processing brands|preferred from brands|indifferent may occur during early processing stages, still in the visual system. However, we found no evidence of a "decision pipeline" that would yield if a brand is preferred or indifferent. The results suggest the existence of a "tagging"-like process in parallel flows in the extrastriate. Network training dynamics aggregate specific processes within the hidden nodes by analyzing the model's hidden layer. This yielded that some nodes contribute to both global brand appraisal and specific brand category classification, shedding light on the neural substrates of decision-making in response to brand stimuli. [ABSTRACT FROM AUTHOR]

Shapley additive explanations are a widely used technique for explaining machine learning models. They can be applied to basically any type of model and provide both global and local explanations. While there are different plots available to visualize Shapley values, there is a lack of suitable visualization for geospatial use cases, resulting in the loss of the geospatial context in traditional plots. This study presents a concept for visualizing Shapley values in geospatial use cases and demonstrate its feasibility through an exemplary use case—predicting bike activity in a rental bike system. The visualizations show that visualizing Shapley values on geographic maps can provide valuable insights that are not visible in traditional plots for Shapley additive explanations. Geovisualizations are recommended for explaining machine learning models in geospatial applications or for extracting knowledge about real-world applications. Suitable visualizations for the considered use case are a proportional symbol map and a mapping of computed Voronoi values to the street network.

This study employs soft computing techniques, including artificial neural network (ANN) models and gene expression programming (GEP), to enhance the prediction of ultimate load in timber pull-out tests under varying environmental conditions. A comprehensive dataset of 202 samples in normal environmental conditions and 324 samples under harsh conditions was gathered. Distinct models were developed for each scenario, achieving commendable accuracies. The ANN performed at 0.91 under normal conditions and 0.99 under harsh conditions, while the GEP performed at 0.91 and 0.94, respectively. The study also predicted free end slip in timber pull-out tests using an ANN model with an accuracy of 0.97. The SHapley values technique was employed to assess the impact of features on the models, revealing specific influential features. In the prediction model for ultimate load under harsh conditions, the rod type was most influential, while under normal conditions, bonded length demonstrated the highest impact. Additionally, the duration of immersion feature had the most substantial effect on predicting free end slip. The final section compared predicted data with experimental values, showing a noteworthy correlation of over 60% between the predicted outputs and established models.

Background: Virtual reality technology has been widely used in surgical simulators, providing new opportunities for assessing and training surgical skills. Machine learning algorithms are commonly used to analyze and evaluate the performance of participants. However, their interpretability limits the personalization of the training for individual participants. Methods: Seventy-nine participants were recruited and divided into three groups based on their skill level in intracranial tumor resection. Data on the use of surgical tools were collected using a surgical simulator. Feature selection was performed using the Minimum Redundancy Maximum Relevance and SVM-RFE algorithms to obtain the final metrics for training the machine learning model. Five machine learning algorithms were trained to predict the skill level, and the support vector machine performed the best, with an accuracy of 92.41% and Area Under Curve value of0.98253. The machine learning model was interpreted using Shapley values to identify the important factors contributing to the skill level of each participant. Results: This study demonstrates the effectiveness of machine learning in differentiating the evaluation and training of virtual reality neurosurgical per- formances. The use of Shapley values enables targeted training by identifying deficiencies in individual skills. Conclusions: This study provides insights into the use of machine learning for personalized training in virtual reality neurosurgery. The interpretability of the machine learning models enables the development of individualized training programs. In addition, this study highlighted the potential of explanatory models in training external skills.

Breast cancer is a critical health issue affecting the well-being of women. Breast cancer is one of the most common causes of the increase in the mortality rate of women around the world. Early detection of breast cancer can, to some extent, decrease the number of deaths caused and also improve treatment outcomes and patient survival rates. Traditional Machine learning (ML) and Deep learning (DL) approaches have proven to be more successful in predicting breast cancer. In terms of privacy and early detection, Federated Learning (FL), a decentralized ML approach, offers a promising solution for training predictive models on distributed healthcare data while ensuring privacy and security. This paper proposes a novel framework that combines the benefits of integrating Shapley values and game theory concepts with FL for breast cancer prediction. The framework uses Shapley values for feature selection from 30 features of the Wisconsin Diagnostic Breast Cancer (WDBC) dataset from University of California Irvine machine learning repository (UCI ML). The framework also addresses the issue of poor-performing clients by introducing a payoff mechanism based on individual client accuracy. Clients with higher accuracy are given greater influence in the model aggregation process, encouraging client competition and improving the overall model performance. Our framework proves to be promising by achieving a prediction accuracy of 94.73% in the FL environment. The proposed approach provides a privacy-preserving solution for breast cancer prediction in an FL environment, by combining Shapley values and game theory. The results of this study can help in the development of more accurate and robust breast cancer prediction models, contributing to improved patient outcomes and healthcare decision-making.

Copyright of Informatica (03505596) is the property of Slovene Society Informatika and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Machine learning (ML) has become more prevalent as a tool used for biogeochemical analysis in agricultural management. However, a common drawback of ML models is the lack of interpretability, as they are black boxes that provide little insight into agricultural management. To overcome this limitation, we compared three tree-based models (decision tree, random forest, and gradient boosting) to explain soil organic matter content through Shapley additive explanations (SHAP). Here, we used nationwide data on field crops, soil, terrain, and climate across South Korea (n = 9584). Using the SHAP method, we identified common primary controls of the models, for example, regions with precipitation levels above 1400 mm and exchangeable potassium levels exceeding 1 cmol+ kg−1, which favor enhanced organic matter in the soil. Different models identified different impacts of macronutrients on the organic matter content in the soil. The SHAP method is practical for assessing whether different ML models yield consistent findings in addressing these inquiries. Increasing the explainability of these models means determining essential variables related to soil organic matter management and understanding their associations for specific instances. [ABSTRACT FROM AUTHOR]

Low‐pressure systems (LPSs) are the primary rainbringers of the South Asian monsoon. Yet, their interactions with the large‐scale monsoon circulation, as well as the highly variable land and sea surfaces they pass over, are complex and generally not well understood. In this article, we present a novel, top‐down approach to investigate these relationships and quantify their importance in describing LPS behaviour. We also show that, if the approach is sufficiently well posed, it is productive at hypothesis generation. For each of five predictands (i.e., LPS intensification rate, propagation speed/direction, post‐landfall survival, peak intensity, and precipitation rate) we train an additive decision‐tree ensemble model using the XGBoost algorithm. Shapley value analysis is then applied to the models to determine which variables are important predictors and to establish their relationship with the predictand, with additional analysis following cases of interest. Novel relationships established using this technique include that LPS vorticity intensifies preferentially in the early morning at the same time as the peak in the diurnal cycle of their convection occurs, that vertical wind shear suppresses continued growth of strong LPSs, that large‐scale barotropic instability plays an important role in both the inland penetration and peak intensity of LPSs, and that LPS propagation depends on the depth of its vortex with shallower LPSs advected by low‐level winds and taller LPSs advected by mid‐level winds. We also use this framework to identify and discuss potential new avenues of research for monsoon LPSs. [ABSTRACT FROM AUTHOR]

Financial inclusion is a fundamental and multidimensional matter that has acquired importance on the global agenda in recent years. In addition, it is still a source of great interest and concern for lawmakers, international organizations, scholars, and financial institutions worldwide. In that regard, this research focuses on Peru to assess the country's financial inclusion condition, which continues to face significant hurdles in providing financial services to its whole population despite economic improvement. The aim of this article is twofold, based on recent data on demand for financial services and financial culture in the country: (1) to empirically test how machine learning methods, such as decision trees, random forests, artificial neural networks, XGBoost, and support vector machines, can be a valuable complement to standard models (i.e., generalized linear models like logistic regression) for assessing financial inclusion in Peru, and (2) to identify the most influential sociodemographic factors on financial inclusion assessment in the country. The results may catalyze the integration of machine learning techniques into the Peruvian financial system, garnering the interest of finance researchers and policymakers committed to augmenting financial access and utilization among Peruvian consumers. [ABSTRACT FROM AUTHOR]

Machine learning (ML) and deep learning (DL) models have been extensively explored for the early diagnosis of various cancer diseases, including Leukemia, with many of them achieving significant performance improvements comparable to those of human experts. However, challenges like limited image data, inaccurate annotations, and prediction reliability still hinder their broad implementation to establish a trustworthy computer-aided diagnosis (CAD) system. This paper introduces a novel weighted-average ensemble model for classifying Acute Lymphoblastic Leukemia, along with a reliable Computer-Aided Diagnosis (CAD) system that combines the strengths of both ML and DL approaches. Initially, a variety of filtering methods are extensively analyzed to determine the most suitable image representation, with subsequent data augmentation techniques to expand the training data. Second, a modified VGG-19 model was proposed with fine-tuning that was utilized as a feature extractor to extract meaningful features from the training samples. Third, A small-loss approach and probabilistic local outlier factor (pLOF) have been developed on the extracted features to address the label noise issue. Fourth, we proposed an weighted-average ensemble model based on the top five models as base learners, with weights calculated based on their model uncertainty to ensure reliable predictions. Fifth, we calculated Shapley values based on cooperative game theory and performed feature selection with different feature combinations to determine the optimal number of features using SHAP. Finally, we integrate these strategies to develop an interpretable CAD system. This system not only predicts the disease but also generates Grad-CAM images to visualize potential affected areas, enhancing both clarity and diagnostic insight. All of our code is provided in the following repository: https://github.com/taareek/leukemia-classification