Explaining a series of models by propagating Shapley values
Ontology highlight
ABSTRACT:
SUBMITTER: Chen H
PROVIDER: S-EPMC9349278 | biostudies-literature | 2022 Jan
REPOSITORIES: biostudies-literature
Similar Datasets
Project description:BackgroundMachine learning (ML) can be an effective tool to extract information from attribute-rich molecular datasets for the generation of molecular diagnostic tests. However, the way in which the resulting scores or classifications are produced from the input data may not be transparent. Algorithmic explainability or interpretability has become a focus of ML research. Shapley values, first introduced in game theory, can provide explanations of the result generated from a specific set of input data by a complex ML algorithm.MethodsFor a multivariate molecular diagnostic test in clinical use (the VeriStrat® test), we calculate and discuss the interpretation of exact Shapley values. We also employ some standard approximation techniques for Shapley value computation (local interpretable model-agnostic explanation (LIME) and Shapley Additive Explanations (SHAP) based methods) and compare the results with exact Shapley values.ResultsExact Shapley values calculated for data collected from a cohort of 256 patients showed that the relative importance of attributes for test classification varied by sample. While all eight features used in the VeriStrat® test contributed equally to classification for some samples, other samples showed more complex patterns of attribute importance for classification generation. Exact Shapley values and Shapley-based interaction metrics were able to provide interpretable classification explanations at the sample or patient level, while patient subgroups could be defined by comparing Shapley value profiles between patients. LIME and SHAP approximation approaches, even those seeking to include correlations between attributes, produced results that were quantitatively and, in some cases qualitatively, different from the exact Shapley values.ConclusionsShapley values can be used to determine the relative importance of input attributes to the result generated by a multivariate molecular diagnostic test for an individual sample or patient. Patient subgroups defined by Shapley value profiles may motivate translational research. However, correlations inherent in molecular data and the typically small ML training sets available for molecular diagnostic test development may cause some approximation methods to produce approximate Shapley values that differ both qualitatively and quantitatively from exact Shapley values. Hence, caution is advised when using approximate methods to evaluate Shapley explanations of the results of molecular diagnostic tests.
| S-EPMC8265031 | biostudies-literature
Project description:The true population-level importance of a variable in a prediction task provides useful knowledge about the underlying data-generating mechanism and can help in deciding which measurements to collect in subsequent experiments. Valid statistical inference on this importance is a key component in understanding the population of interest. We present a computationally efficient procedure for estimating and obtaining valid statistical inference on the Shapley Population Variable Importance Measure (SPVIM). Although the computational complexity of the true SPVIM scales exponentially with the number of variables, we propose an estimator based on randomly sampling only Θ(n) feature subsets given n observations. We prove that our estimator converges at an asymptotically optimal rate. Moreover, by deriving the asymptotic distribution of our estimator, we construct valid confidence intervals and hypothesis tests. Our procedure has good finite-sample performance in simulations, and for an in-hospital mortality prediction task produces similar variable importance estimates when different machine learning algorithms are applied.
| S-EPMC8057672 | biostudies-literature
Project description:Neurological disorders such as Autism Spectrum Disorder (ASD), Schizophrenia (SCH), Bipolar Disorder (BD), and Major Depressive Disorder (MDD) affect millions of people worldwide, yet their molecular mechanisms remain poorly understood. This study describes the application of the Comparative Analysis of Shapley values (CASh) to transcriptomic data from nine datasets associated with these complex disorders, demonstrating its effectiveness in identifying differentially expressed genes (DEGs). CASh, which combines Game Theory with Bootstrap resampling, offers a robust alternative to traditional statistical methods by assessing the contribution of each gene in the broader context of the complete dataset. Unlike conventional approaches, CASh is highly effective at detecting subtle but meaningful molecular patterns that are often missed. These findings highlight the potential of CASh to enhance the precision of transcriptomic analysis, providing a deeper understanding of the molecular mechanisms underlying these disorders and establishing a solid basis to improve diagnostic techniques and developing more targeted therapeutic interventions.
| S-EPMC11726880 | biostudies-literature
Project description:In this paper we apply a series of Machine Learning models to a recently published unique dataset on the mortality of COVID19 patients. We use a dataset consisting of blood samples of 375 patients admitted to a hospital in the region of Wuhan, China. There are 201 patients who survived hospitalisation and 174 patients who died whilst in hospital. The focus of the paper is not only on seeing which Machine Learning model is able to obtain the absolute highest accuracy but more on the interpretation of what the Machine Learning models provides. We find that age, days in hospital, Lymphocyte and Neutrophils are important and robust predictors when predicting a patients mortality. Furthermore, the algorithms we use allows us to observe the marginal impact of each variable on a case-by-case patient level, which might help practicioneers to easily detect anomalous patterns. This paper analyses the global and local interpretation of the Machine Learning models on patients with COVID19.
| S-EPMC7948528 | biostudies-literature
Project description:BackgroundTo enable automatic searches, alignments, and model combination, the elements of systems biology models need to be compared and matched across models. Elements can be identified by machine-readable biological annotations, but assigning such annotations and matching non-annotated elements is tedious work and calls for automation.ResultsA new method called "semantic propagation" allows the comparison of model elements based not only on their own annotations, but also on annotations of surrounding elements in the network. One may either propagate feature vectors, describing the annotations of individual elements, or quantitative similarities between elements from different models. Based on semantic propagation, we align partially annotated models and find annotations for non-annotated model elements.ConclusionsSemantic propagation and model alignment are included in the open-source library semanticSBML, available on sourceforge. Online services for model alignment and for annotation prediction can be used at http://www.semanticsbml.org.
| S-EPMC3340317 | biostudies-literature
Project description:The support vector machine (SVM) algorithm is popular in chemistry and drug discovery. SVM models have black box character. Their predictions can be interpreted through feature weighting or the model-agnostic Shapley additive explanations (SHAP) formalism that locally approximates Shapley values (SVs) originating from game theory. We introduce an algorithm termed SV-expressed Tanimoto similarity (SVETA) for the exact calculation of SVs to explain SVM models employing the Tanimoto kernel, the gold standard for the assessment of molecular similarity. For a model system, the exact calculation of SVs is demonstrated. In an SVM-based compound classification task from drug discovery, only a limited correlation between exact SV and SHAP values is observed, prohibiting the use of approximate values for rationalizing predictions. For exemplary test compounds, atom-based mapping of prioritized features delineates coherent substructures that closely resemble those obtained by analyzing independently derived random forest models, thus providing consistent explanations.
| S-EPMC9464958 | biostudies-literature
Project description:Amyotrophic lateral sclerosis (ALS) is a prototypical neurodegenerative disease characterized by progressive degeneration of motor neurons to severely effect the functionality to control voluntary muscle movement. Most of the non-additive genetic aberrations responsible for ALS make its molecular classification very challenging along with limited sample size, curse of dimensionality, class imbalance and noise in the data. Deep learning methods have been successful in many other related areas but have low minority class accuracy and suffer from the lack of explainability when used directly with RNA expression features for ALS molecular classification. In this paper, we propose a deep-learning-based molecular ALS classification and interpretation framework. Our framework is based on training a convolution neural network (CNN) on images obtained from converting RNA expression values into pixels based on DeepInsight similarity technique. Then, we employed Shapley additive explanations (SHAP) to extract pixels with higher relevance to ALS classifications. These pixels were mapped back to the genes which made them up. This enabled us to classify ALS samples with high accuracy for a minority class along with identifying genes that might be playing an important role in ALS molecular classifications. Taken together with RNA expression images classified with CNN, our preliminary analysis of the genes identified by SHAP interpretation demonstrate the value of utilizing Machine Learning to perform molecular classification of ALS and uncover disease-associated genes.
| S-EPMC8626003 | biostudies-literature
Project description:BackgroundIn the field of implementation research, there is an increased interest in use of theory when designing implementation research studies involving behavior change. In 2003, we initiated a series of five studies to establish a scientific rationale for interventions to translate research findings into clinical practice by exploring the performance of a number of different, commonly used, overlapping behavioral theories and models. We reflect on the strengths and weaknesses of the methods, the performance of the theories, and consider where these methods sit alongside the range of methods for studying healthcare professional behavior change.MethodsThese were five studies of the theory-based cognitions and clinical behaviors (taking dental radiographs, performing dental restorations, placing fissure sealants, managing upper respiratory tract infections without prescribing antibiotics, managing low back pain without ordering lumbar spine x-rays) of random samples of primary care dentists and physicians. Measures were derived for the explanatory theoretical constructs in the Theory of Planned Behavior (TPB), Social Cognitive Theory (SCT), and Illness Representations specified by the Common Sense Self Regulation Model (CSSRM). We constructed self-report measures of two constructs from Learning Theory (LT), a measure of Implementation Intentions (II), and the Precaution Adoption Process. We collected data on theory-based cognitions (explanatory measures) and two interim outcome measures (stated behavioral intention and simulated behavior) by postal questionnaire survey during the 12-month period to which objective measures of behavior (collected from routine administrative sources) were related. Planned analyses explored the predictive value of theories in explaining variance in intention, behavioral simulation and behavior.ResultsResponse rates across the five surveys ranged from 21% to 48%; we achieved the target sample size for three of the five surveys. For the predictor variables, the mean construct scores were above the mid-point on the scale with median values across the five behaviors generally being above four out of seven and the range being from 1.53 to 6.01. Across all of the theories, the highest proportion of the variance explained was always for intention and the lowest was for behavior. The Knowledge-Attitudes-Behavior Model performed poorly across all behaviors and dependent variables; CSSRM also performed poorly. For TPB, SCT, II, and LT across the five behaviors, we predicted median R2 of 25% to 42.6% for intention, 6.2% to 16% for behavioral simulation, and 2.4% to 6.3% for behavior.ConclusionsWe operationalized multiple theories measuring across five behaviors. Continuing challenges that emerge from our work are: better specification of behaviors, better operationalization of theories; how best to appropriately extend the range of theories; further assessment of the value of theories in different settings and groups; exploring the implications of these methods for the management of chronic diseases; and moving to experimental designs to allow an understanding of behavior change.
| S-EPMC3500222 | biostudies-literature
Project description:The emergence of deep learning networks raises a need for explainable AI so that users and domain experts can be confident applying them to high-risk decisions. In this paper, we leverage data from the latent space induced by deep learning models to learn stereotypical representations or "prototypes" during training to elucidate the algorithmic decision-making process. We study how leveraging prototypes effect classification decisions of two dimensional time-series data in a few different settings: (1) electrocardiogram (ECG) waveforms to detect clinical bradycardia, a slowing of heart rate, in preterm infants, (2) respiration waveforms to detect apnea of prematurity, and (3) audio waveforms to classify spoken digits. We improve upon existing models by optimizing for increased prototype diversity and robustness, visualize how these prototypes in the latent space are used by the model to distinguish classes, and show that prototypes are capable of learning features on two dimensional time-series data to produce explainable insights during classification tasks. We show that the prototypes are capable of learning real-world features - bradycardia in ECG, apnea in respiration, and articulation in speech - as well as features within sub-classes. Our novel work leverages learned prototypical framework on two dimensional time-series data to produce explainable insights during classification tasks.
| S-EPMC8050893 | biostudies-literature
Project description:Identification of those at greatest risk of death due to the substantial threat of COVID-19 can benefit from novel approaches to epidemiology that leverage large datasets and complex machine-learning models, provide data-driven intelligence, and guide decisions such as intensive-care unit admission (ICUA). The objective of this study is two-fold, one substantive and one methodological: substantively to evaluate the association of demographic and health records with two related, yet different, outcomes of severe COVID-19 (viz., death and ICUA); methodologically to compare interpretations based on logistic regression and on gradient-boosted decision tree (GBDT) predictions interpreted by means of the Shapley impacts of covariates. Very different association of some factors, e.g., obesity and chronic respiratory diseases, with death and ICUA may guide review of practice. Shapley explanation of GBDTs identified varying effects of some factors among patients, thus emphasising the importance of individual patient assessment. The results of this study are also relevant for the evaluation of complex automated clinical decision systems, which should optimise prediction scores whilst remaining interpretable to clinicians and mitigating potential biases.
| S-EPMC8259985 | biostudies-literature