关键词： Random forest classifier   TCR-pMHC complexes   Binding affinity  

摘要： The activation of T cells is triggered by the interactions of T cell receptors (TCRs) with their epitopes, which are peptides presented by major histocompatibility complex (MHC) on the surfaces of antigen presenting cells (APC). While each TCR can only recognize a specific subset from a large repertoire of peptide-MHC (pMHC) complexes, it is very often that peptides in this subset share little sequence similarity. This is known as the specificity and cross-reactivity of T cells, respectively. The binding affinities between different types of TCRs and pMHC are the major driving force to shape this specificity and cross-reactivity in T cell recognition. The binding affinities, furthermore, are determined by the sequence and structural properties at the interfaces between TCRs and pMHC. Fortunately, a wealth of data on binding and structures of TCR-pMHC interactions becomes publicly accessible in online resources, which offers us the opportunity to develop a random forest classifier for predicting the binding affinities between TCR and pMHC based on the structure of their complexes. Specifically, the structure and sequence of a given complex were projected onto a high-dimensional feature space as the input of the classifier, which was then trained by a large-scale benchmark dataset. Based on the cross-validation results, we found that our machine learning model can predict if the binding affinity of a given TCR-pMHC complex is stronger or weaker than a predefined threshold with an overall accuracy approximately around 75 %. The significance of our prediction was estimated by statistical analysis. Moreover, more than 60 % of binding affinities in the ATLAS database can be successfully classified into groups within the range of 2 kcal/mol. Additionally, we show that TCR-pMHC complexes with strong binding affinity prefer hydrophobic interactions between amino acids with large aromatic rings instead of electrostatic interactions. Our results therefore provide insights to d

关键词： Interpretable machine learning   Matrix factorization   Graph model   Homotrimer  

摘要： Protein-protein interaction plays an important role in life activities. A more fine-grained analysis, such as residues and atoms level, will better benefit us to understand the mechanism for inter-protein interaction and drug design. The development of efficient computational methods to reduce trials and errors, as well as assisting experimental researchers to determine the complex structure are some of the ongoing studies in the field. The research of trimer protein interface, especially homotrimer, has been rarely studied. In this paper, we proposed an interpretable machine learning method for homo-trimeric protein interface residue pairs prediction. The structure, sequence, and physicochemical information are intergraded as feature input fed to model for training. Graph model is utilized to present spatial information for intra-protein. Matrix factorization captures the different features' interactions. Kernel function is designed to auto-acquire the adjacent information of our target residue pairs. The accuracy rate achieves 54.5% in an independent test set. Sequence and structure alignment exhibit the ability of model self-study. Our model indicates the biological significance between sequence and structure, and could be auxiliary for reducing trials and errors in the fields of protein complex determination and protein-protein docking, etc. Significance: Protein complex structures are significant for understanding protein function and promising functional protein design. With data increasing, some computational tools have been developed for protein complex residue contact prediction, which is one of the most significant steps for complex structure prediction. But for homo-trimeric protein, the sequence-based deep learning predictors are infeasible for homologous sequences, and the algorithm black box prevents us from understanding of each step operation. In this way, we propose an interpreting machine learning method for homo-trimeric protein interface residue-

关键词： Machine learning   Computational materials   Ferroelectrics   Solid Solutions  

摘要： In this work, we studied the application of machine learning for the prediction of the Curie temperature (T-c) of ferroelectric BiMe'Me"O-3-PbTiO3 systems (where Me' and Me" are metal cations). We found that among the studied K-nearest neighbor (KNN), support vector regression (SVR) and random forest (RF) methods, RF obtains the best performance and is insensitive to the choice of hyperparameters. SVR results show a strong sensitivity to the choice of hyperparameters and obtained T-c predictions with significantly lower accuracy than RF even after hyperparameter optimization. KNN results show poor accuracy and are essentially unusable with an incomplete feature set and are only qualitatively accurate with a complete feature set. With regard to the choice of features for accurate prediction of the Ferroelectric (FE) systems, we find that Bi content and B-cation valence, ionic radius and ionic displacements form the irreducible set of features such that these features or their equivalents must be used to obtain quantitatively accurate T-c predictions. We also find that homovalent and hetemvalent BiMe'Me '' O-3-PbTiO3 solid solutions form distinct classes of compounds with different behaviors so that both types must be included in the input data set to obtain high predictive accuracy. Our work confirms that for the small data sets typically available in materials science, careful selection of the input system, features and ML methods is required to enable accurate model construction and discovery of previously unknown relationships, but can be achieved in a systematic manner.

关键词： Cost sensitivity   Hepatic encephalopathy   Disease risk prediction   Weighted random forest   Weighted support vector machine  

摘要： Objective: Hepatic encephalopathy (HE) is among the most common complications of cirrhosis. Data for cirrhosis with HE is typically unbalanced. Traditional statistical methods and machine learning algorithms thus cannot identify a few classes. In this paper, we use machine learning algorithms to construct a risk prediction model for liver cirrhosis complicated by HE to improve the efficiency of its prediction. Method: We collected medical data from 1,256 patients with cirrhosis and performed preprocessing to extract 81 features from these irregular data. To predict HE in cirrhotic patients, we compared several classification methods: logistic regression, weighted random forest (WRF), SVM, and weighted SVM (WSVM). We also used an additional 722 patients with cirrhosis for external validation of the model. Results: The WRF, WSVM, and logistic regression models exhibited better recognition ability for patients with HE than traditional machine learning models (sensitivity > 0.70), but their ability to identify patients with uncomplicated HE was slightly lower (specificity approximately 85%). The comprehensive evaluation index of the traditional model was higher than those of other models (G-means > 0.80 and F-measure > 0.40). For the WRF, the G-means (0.82), F-measure (0.46), and AUC (0.82) were superior to those of the logistic regression and WSVM models, which means that it can better predict the incidence of HE in patients. Conclusion: The WRF model is more suitable for the classification of unbalanced medical data and can be used to construct a risk prediction and evaluation system for liver cirrhosis complicated with HE. The probabilistic prediction models of WRF can help clinicians identify high-risk patients with HE. (c) 2021 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license ( http://***/licenses/by-nc-nd/4.0/ )

关键词： machine learning   ensemble-based machine learning algorithm   bagging random forest algorithm   structural health monitoring   guided waves   Lamb waves   structural damage estimation  

摘要： Lamb wave-based damage estimation have great potential for structural health monitoring. However, designing a generalizable model that predicts accurate and reliable damage quantification result is still a practice challenge due to complex behavior of waves with different damage severities. In the recent years, machine learning (ML) algorithms have been proven to be an efficient tool to analyze damage-modulated Lamb wave signals. In this study, ensemble-based ML algorithms are employed to develop a generalizable crack quantification model for thin metallic plates. For this, the scattering of Lamb wave signals due to different configuration of crack dimension and orientation is extensively studied. Various finite element simulations signals, representing distinct crack severities in term of crack length, penetration and orientation are acquired. Realizing that both temporal and spectral information of signal is extremely important to damage quantification, three time-frequency (TF) based damage sensitive indices namely energy concentration, TF flux and coefficient of energy variance are proposed. These damage features are extracted by employing smoothed-pseudo Wigner-Ville distribution. After that data augmentation technique based on the spline-based interpolation is applied to enhance the size of the dataset. Eventually, these fully developed damage dataset is deployed to train ensemble-based models. Here we propose separate model network, in which different models are trained and then link together to predict new and unseen datasets. The performance of the proposed model is demonstrated by two cases: first simulated data incorporated with high artificial noises are employed to test the model and in the second scenario, experimental data in raw form are used. Results indicate that the proposed model has the potential to develop a general model that yields reliable answer for crack quantification.

关键词： Incident hypertension   Genetic factor   Risk model   Predictive performance  

摘要： Current studies have shown the controversial effect of genetic risk scores (GRSs) in hypertension prediction. Machine learning methods are used extensively in the medical field but rarely in the mining of genetic information. This study aims to determine whether genetic information can improve the prediction of incident hypertension using machine learning approaches in a prospective study. The study recruited 4592 subjects without hypertension at baseline from a cohort study conducted in rural China. A polygenic risk score (PGGRS) was calculated using 13 SNPs. According to a ratio of 7:3, subjects were randomly allocated to the train and test datasets. Models with and without the PGGRS were established using the train dataset with Cox regression, artificial neural network (ANN), random forest (RF), and gradient boosting machine (GBM) methods. The discrimination and reclassification of models were estimated using the test dataset. The PGGRS showed a significant association with the risk of incident hypertension (HR (95% CI), 1.046 (1.004, 1.090), P = 0.031) irrespective of baseline blood pressure. Models that did not include the PGGRS achieved AUCs (95% CI) of 0.785 (0.763, 0.807), 0.790 (0.768, 0.811), 0.838 (0.817, 0.857), and 0.854 (0.835, 0.873) for the Cox, ANN, RF, and GBM methods, respectively. The addition of the PGGRS led to the improvement of the AUC by 0.001, 0.008, 0.023, and 0.017;IDI by 1.39%, 2.86%, 4.73%, and 4.68%;and NRI by 25.05%, 13.01%, 44.87%, and 22.94%, respectively. Incident hypertension risk was better predicted by the traditional+PGGRS model, especially when machine learning approaches were used, suggesting that genetic information may have the potential to identify new hypertension cases using machine learning methods in resource-limited areas.

关键词： electronic nose   gas sensor array   machine learning   neural networks   review  

摘要： Machine learning methods enable the electronic nose (E-Nose) for precise odor identification with both qualitative and quantitative analysis. Advanced machine learning methods are crucial for the E-Nose to gain high performance and strengthen its capability in many applications, including robotics, food engineering, environment monitoring, and medical diagnosis. Recently, many machine learning techniques have been studied, developed, and integrated into feature extraction, modeling, and gas sensor drift compensation. The purpose of feature extraction is to keep robust pattern information in raw signals while removing redundancy and noise. With the extracted feature, a proper modeling method can effectively use the information for prediction. In addition, drift compensation is adopted to relieve the model accuracy degradation due to the gas sensor drifting. These recent advances have significantly promoted the prediction accuracy and stability of the E-Nose. This review is engaged to provide a summary of recent progress in advanced machine learning methods in E-Nose technologies and give an insight into new research directions in feature extraction, modeling, and sensor drift compensation.

关键词： Field effect transistors   Optimization   Logic gates   Nanoscale devices   Effective mass   Performance evaluation   Capacitance   2-D field-effect transistors (FETs)   active learning   global optimization   machine learning (ML)   multiobjective optimization (MOO)   nanoscale transistors   transistor simulation  

摘要： Design optimization of emerging nanoscale transistor technologies often requires careful design tradeoff between many objectives, including speed, power, variability, and so on. By leveraging machine learning (ML) methods, we develop a multiobjective optimization (MOO) framework for 2-D-material-based field-effect transistors (FETs) near the scaling limit. The MOO design framework performs gradient-free efficient global optimization and offers the option of using active learning. Optimum designs with a tradeoff between transistor speed, power, and variability are identified automatically for transition metal dichalcogenide (TMDC) and black phosphorene FETs by applying the MOO design framework that couples ML methods to quantum transport device simulations. The design optimization results show that the International Roadmap of Devices and Systems (IRDS) target of 2025 and 2028 technology nodes can be met by 2-D FETs.

关键词： Machine learning   Genetic algorithm   Worker disability   Occupational accident   Construction safety   Safety management   Tree-based ensemble models   Artificial intelligence  

摘要： The construction industry is among the riskiest industries around the world. Hence, the preliminary studies exploring the consequences of occupational accidents have received considerable attention in research society. This study aims to develop a comprehensive framework to predict the post-accident disability status of construction workers. The dataset comprising 47,938 construction accidents recorded in Turkey was subjected to a detailed multi-step feature engineering approach, including data encoding, data scaling, dimension reduction, and data resampling. Predictions were performed through four tree-based ensemble machine learning models: Random Forest, XGBoost, AdaBoost, and Extra Trees, as well as a state-of-the-art optimization method for hyperparameter tuning, Genetic Algorithm (GA). GA-XGBoost presented the highest prediction rate with 0.8292 in terms of accuracy and 0.8120 with respect to AUROC. The findings may aid in predicting construction workers' post-accident disability status, resulting in a safer working environment and productivity planning in construction projects.

关键词： Remote sensing   Spectroscopy   Satellite   Unmanned Aerial Vehicles   Partial least squares regression   Random forests   Multiple linear regression  

摘要： The study of grasslands using machine learning (ML) methods combined with proximal/remote sensing data (RS) has been steadily increasing in the last decades. Available algorithms range from a primarily academic use to more widespread practical applications intended at helping farm management. Here, we review the use of ML methods applied to aboveground biomass (AGB) estimation in grassland systems. Based on 26 recent papers, we perform a literature review of the topic to identify common practices, namely the relation between estimation performance and the ML method used, data sources, and scale (local/regional). In order to identify the relation between the characteristics of the studies and the estimation accuracy, we use descriptive and correlation analysis. In spite of a surge in the number of papers and application examples, there is no evidence that the estimation performance of the algorithms has been improving over time. In all approaches used by the authors of the papers herein considered, the number of field samples, RS data source, and species composition of the grassland systems are the most relevant variables to explain the estimation accuracy. This accuracy increases with the number of field samples until it plateaus, hinting at the existence of an optimum level for monitoring efforts. Accuracy also increases with the proximity of the sensor to the field, i.e., on average accuracy is higher using field spectroscopy than using satellite data. There is no evidence that any particular ML method is more suited to this problem. The literature also displays significant limitations in terms of its applications of the ML algorithms. For example, a limited number of papers validated the models, casting doubt on the potential of the models for generalized application. Despite those limitations, and considering the advancements verified, we expect that, in the near future, ML methods combined with RS/proximal data will continue to improve and be helpful for farm m