Recursive Feature Elimination Method Research Articles

A novel enhanced decision tree model for detecting chronic kidney disease

Prediction of diseases is sensitive as any error can result in the wrong person's treatment or not treating the right patient. Besides, some features distinguish a disease from curable to fatal or curable to chronic disease. Data mining techniques have been widely used in health-related research. The researchers, so far, could attain around 97 percent accuracy using several methods. Some researchers have demonstrated that the selection of correct features increases the prediction accuracy. This research work propose a method to distinguish between chronic and non-chronic kidney disease, identify its crucial features without reducing the accuracy of prediction, and a prediction algorithm to eliminate the possibility of under or overfitting. This study uses the recursive feature elimination (RFE) method that selects an optimal subset of features and an ensemble algorithm, the enhanced decision tree (EDT), to predict the disease. The results obtained in this paper show that the accuracy level of EDT is not changed with the removal of less significant features, thus enabling the decision-makers to concentrate on few features to reduce time and error of treatment. EDT establishes substantially high consistency in predicting, with or without feature selection, the disease.

Prediction of Type 2 Diabetes Based on Machine Learning Algorithm.

Prediction of type 2 diabetes (T2D) occurrence allows a person at risk to take actions that can prevent onset or delay the progression of the disease. In this study, we developed a machine learning (ML) model to predict T2D occurrence in the following year (Y + 1) using variables in the current year (Y). The dataset for this study was collected at a private medical institute as electronic health records from 2013 to 2018. To construct the prediction model, key features were first selected using ANOVA tests, chi-squared tests, and recursive feature elimination methods. The resultant features were fasting plasma glucose (FPG), HbA1c, triglycerides, BMI, gamma-GTP, age, uric acid, sex, smoking, drinking, physical activity, and family history. We then employed logistic regression, random forest, support vector machine, XGBoost, and ensemble machine learning algorithms based on these variables to predict the outcome as normal (non-diabetic), prediabetes, or diabetes. Based on the experimental results, the performance of the prediction model proved to be reasonably good at forecasting the occurrence of T2D in the Korean population. The model can provide clinicians and patients with valuable predictive information on the likelihood of developing T2D. The cross-validation (CV) results showed that the ensemble models had a superior performance to that of the single models. The CV performance of the prediction models was improved by incorporating more medical history from the dataset.

Towards a novel set of GPS-derived metrics to identify the differences between mobility patterns of cognitively intact older adults and older adults with dementia

<h3>Introduction</h3> Maintaining an active lifestyle and participating in social activities are key components of healthy ageing. These components rely on the individuals' ability to remain mobile out of home. The global positioning system (GPS) is increasingly used to assess outdoor mobility of older adults. However, there is a gap in establishing a framework that recognizes the differences between mobility patterns of cognitively intact older adults and older adults with dementia (OAwD). We propose a set of GPS-derived metrics to identify these differences, which can be used to evaluate changes in mobility patterns of OAwD over time. <h3>Methods</h3> We analyzed the mobility profiles of 15 older adults from the greater Toronto area using GPS technology. Participants were aged between 65 and 90 years (M=74.9, SD=7.05) and were either cognitively healthy (n = 8) or diagnosed with dementia (n = 7). First, we created GPS-based features representing different dimensions of outdoor mobility. These features include (1) the typical distance covered by individuals, (2) randomness of the travel patterns, (3) the spatial variation of the GPS locations, (4) the number of distinct stops visited per day, (5) the number of daily trips away from home, (6) the score of three levels of outdoor life-space, and (7) the maximum distance travelled from home. Then, we used Recursive Feature Elimination (RFE) with Random Forest (RF) to select the most important mobility features with respect to cognitive status. <h3>Results</h3> A total of 55,580 GPS points were collected by the 15 unique participants over a period of 4 to 8 weeks. The RFE method indicated that features (1), (2), (3), and (5) had the strongest association with cognitive status. Our results suggested that OAwD displayed more predictability (i.e. smaller randomness) in travel patterns compared to the controls (OAwD: M=4.82, SD=0.57 vs. CTL:M=5.62, SD=0.71). Furthermore, the OAwD made fewer daily out-of-home trips compared to the controls (OAwD: M=1.06, SD=0.79 vs. CTL: M=1.65, SD=0.45), and covered smaller distances compared to the controls (OAwD: M=12.27, SD=11.73 km vs. CTL: M=15.78, SD=24.31 km). <h3>Conclusions</h3> The proposed set of GPS-based features identified the differences between mobility patterns of cognitively intact older adults and OAwD. <h3>Funding</h3> Canadian Consortium on Neurodegeneration in Aging

MGRFE: Multilayer Recursive Feature Elimination Based on an Embedded Genetic Algorithm for Cancer Classification.

Microarray gene expression data have become a topic of great interest for cancer classification and for further research in the field of bioinformatics. Nonetheless, due to the "large p, small n" paradigm of limited biosamples and high-dimensional data, gene selection is becoming a demanding task, which is aimed at selecting a minimal number of discriminatory genes associated closely with a phenotype. Feature or gene selection is still a challenging problem owing to its nondeterministic polynomial time complexity and thus most of the existing feature selection algorithms utilize heuristic rules. A multilayer recursive feature elimination method based on an embedded integer-coded genetic algorithm, MGRFE, is proposed here, which is aimed at selecting the gene combination with minimal size and maximal information. On the basis of 19 benchmark microarray datasets including multiclass and imbalanced datasets, MGRFE outperforms state-of-the-art feature selection algorithms with better cancer classification accuracy and a smaller selected gene number. MGRFE could be regarded as a promising feature selection method for high-dimensional datasets especially gene expression data. Moreover, the genes selected by MGRFE have close biological relevance to cancer phenotypes. The source code of our proposed algorithm and all the 19 datasets used in this paper are available at https://github.com/Pengeace/MGRFE-GaRFE.

Estimation of forest aboveground biomass using combination of Landsat 8 and Sentinel-1A data with random forest regression algorithm in Himalayan Foothills

Forest aboveground biomass (AGB) plays an indispensable role in the terrestrial carbon cycle and its dynamics. It also provides baseline data for developing sustainable management strategies in the region. In the present study, a decision-tree based random forest (RF) algorithm was used to estimate AGB for the different forest types in Doon valley, situated in the Himalayan foothills of India. Fifty-one spectral and textural variables were initially extracted using Landsat 8 Operational Land Imager and Sentinel-1A, which were further reduced to twenty optimum variables using the recursive feature elimination (RFE) method. These optimum variables were finally used to map AGB. Results showed that the spatial distribution of AGB ranged from 46.36 to 596.15 Mg ha−1 with good correlation (R2 = 0.87, RMSEr = 18.7%, RMSE = 62.56 Mg ha−1) between the observed and predicted AGB. This study validated the synergistic use of remote sensing, field data, and RF algorithm to precisely predict the spatial distribution of AGB.

Improving satellite retrieval of oceanic particulate organic carbon concentrations using machine learning methods

Particulate organic carbon (POC) plays vital roles in marine carbon cycle, serving as a part of “biological pump” moving carbon to the deep ocean. The blue-to-green band ratio algorithm is applied operationally to derive POC concentrations in global oceans; it, however, tends to underestimate high values in optically complex waters. With an attempt to develop accurate and robust oceanic POC models, this study aimed to explore machine learning methods in satellite retrieval of POC concentrations. Three machine learning methods, i.e. extreme gradient boosting (XGBoost), support vector machine (SVM) and artificial neural network (ANN), were tested, and the recursive feature elimination (RFE) method was employed to identify sensitive features. Matchups of global in situ POC measurements and Ocean Colour Climate Change Initiative (OC-CCI) products were used to train and evaluate POC models. Results showed that machine learning methods produced obvious better performance than the blue-to-green band ratio algorithm, and XGBoost was the most robust among the tested three machine learning methods. However, the blue-to-green band ratio algorithm still worked well for clear open ocean waters with low POC, and ANN was more effective for optically complex waters with extremely high POC. This study provided globally applicable methods for satellite retrieval of POC concentrations, which should be helpful for studying POC dynamics in global oceans as well as in productive marginal seas.

Random forest of perfect trees: concept, performance, applications and perspectives.

MotivationThe principle of Breiman's random forest (RF) is to build and assemble complementary classification trees in a way that maximizes their variability. We propose a new type of random forest that disobeys Breiman’s principles and involves building trees with no classification errors in very large quantities. We used a new type of decision tree that uses a neuron at each node as well as an in-innovative half Christmas tree structure. With these new RFs, we developed a score, based on a family of ten new statistical information criteria, called Nguyen information criteria (NICs), to evaluate the predictive qualities of features in three dimensions.ResultsThe first NIC allowed the Akaike information criterion to be minimized more quickly than data obtained with the Gini index when the features were introduced in a logistic regression model. The selected features based on the NICScore showed a slight advantage compared to the support vector machines—recursive feature elimination (SVM-RFE) method. We demonstrate that the inclusion of artificial neurons in tree nodes allows a large number of classifiers in the same node to be taken into account simultaneously and results in perfect trees without classification errors.Availability and implementationThe methods used to build the perfect trees in this article were implemented in the ‘ROP’ R package, archived at https://cran.r-project.org/web/packages/ROP/index.html.Supplementary informationSupplementary data are available at Bioinformatics online.

A hybrid optimization method of factor screening predicated on GeoDetector and Random Forest for Landslide Susceptibility Mapping

The aim of this study was to develop a hybrid model (Geo-RFE-RF) for Landslide Susceptibility Mapping (LSM) predicated on GeoDetector and Random Forest (RF) using the Recursive Feature Elimination (RFE-RF) method for eliminating redundant and noise factors. At the outset, for the sample of 1522 investigated landslides and 1522 non-landslides, twenty-two factors were chosen as the initial landslide-conditioning factors to construct a spatial database. Subsequently, the GeoDetector and RFE-RF methods were adopted to eliminate the least effective factors, respectively, with the Geo-RFE-RF model being formulated with the combination factors of these two methods. Finally, the performance of the Geo-RFE-RF and RF model with twenty-two initial factors (In-RF) were compared and assessed, and the higher accuracy model was employed to generate a LSM in a case study area, Fengjie County (China). The results indicate that, the Area Under Curve, Accuracy, Precision, and F1 Score of the test dataset is increased by 0.9%, 0.4%, 1.5%, and 0.3%, respectively, under the Geo-RFE-RF model, as compared to the In-RF model. The conditioning factors used to construct the model have been reduced from twenty-two to thirteen, but the predictive ability of the Geo-RFE-RF model performs better, proving the effectiveness of the hybrid model that combines the factors from GeoDetector and RFE-RF methods. This hybrid model not only considers the spatial pattern characteristics of spatial data for screening factors, but the selected factors are also in good agreement with the adopted machine learning method, offering potential use as a general framework for eliminating redundant and noise factors in LSM across the globe.

Pulmonary MRI Radiomics and Machine Learning: Effect of Intralesional Heterogeneity on Classification of Lesion

To investigate the effect of intralesional heterogeneity on differentiating benign and malignant pulmonary lesions, quantitative magnetic resonance imaging (MRI) radiomics, and machine learning methods were adopted.A total of 176 patients with multiparametric MRI were involved in this exploratory study. To investigate the effect of intralesional heterogeneity on lesion classification, a radiomics model called tumor heterogeneity model was developed and compared to the conventional radiomics model based on the entire tumor. In tumor heterogeneity model, each lesion was divided into five sublesions depending on the spatial location through clustering algorithm. From the five sublesions in multi-parametric MRI sequences, 1100 radiomics features were extracted. The recursive feature elimination method was employed to select features and support vector machine classifier was used to distinguish benign and malignant lesion. The performance of classification was evaluated with the receiver operating characteristic curve and the area under the curve (AUC) was the figure of merit. The 3-fold cross-validation (CV) with and without nesting was used to validate the model, respectively.The tumor heterogeneity model (AUC = 0.74 ± 0.04 and 0.90 ± 0.03, CV with and without nesting, respectively) outperforms conventional model (AUC = 0.68 ± 0.04 and 0.87 ± 0.03). The difference between the two models is statistically significant (p = 0.03) for lesions greater than 18.80 cm3.Intralesional heterogeneity influences the classification of pulmonary lesions. The tumor heterogeneity model tends to perform better than conventional radiomics model.

Prediction of heat waves over Pakistan using support vector machine algorithm in the context of climate change

Many efficient forecasting models have been found to fail or show low skill due to the changes in the predictor–predictand relationship with the changes in global climate. An attempt has been taken to develop a climate change resilient heatwave prediction model using machine learning (ML) algorithms known as Support Vector Machines (SVM), random forest and artificial neural network. The National Centres for Environmental Prediction/National Centre for Atmospheric Research reanalysis data of ocean-atmospheric variables were used as the predictors of ML models for forecasting the number of heatwave days (HWDs) in the summer of Pakistan. An SVM based recursive feature elimination method was used to select the skilful predictors. The ML models were developed by considering a moving window of 29 years with a time step of 5 years to incorporate the changes in the relation of HWDs with its predictors due to climate change. The result showed changes in the relationship of HWDs with all the ocean-atmospheric variables considered in this study as probable predictors, which indicates the necessity of forward-rolling approach proposed in this study for the development of climate change resilient forecasting model. The relative performance of ML showed the higher capability of SVM to predict HWDs with an %NRMSE of 36, R2 of 0.87, md score of 0.76 and an rSD of 0.88 during the validation period. The result revealed the potential of SVM model to be used for reliable forecasting of heatwaves in the context of climate change.

A Machine Learning Approach to EEG-based Prediction of Human Affective States Using Recursive Feature Elimination Method

Emotion recognition, as a branch of affective computing, has attracted great attention in the last decades as it can enable more natural brain-computer interface systems. Electroencephalography (EEG) has proven to be an effective modality for emotion recognition, with which user affective states can be tracked and recorded, especially for primitive emotional events such as arousal and valence. Although brain signals have been shown to correlate with emotional states, the effectiveness of proposed models is somewhat limited. The challenge is improving accuracy, while appropriate extraction of valuable features might be a key to success. This study proposes a framework based on incorporating fractal dimension features and recursive feature elimination approach to enhance the accuracy of EEG-based emotion recognition. The fractal dimension and spectrum-based features to be extracted and used for more accurate emotional state recognition. Recursive Feature Elimination will be used as a feature selection method, whereas the classification of emotions will be performed by the Support Vector Machine (SVM) algorithm. The proposed framework will be tested with a widely used public database, and results are expected to demonstrate higher accuracy and robustness compared to other studies. The contributions of this study are primarily about the improvement of the EEG-based emotion classification accuracy. There is a potential restriction of how generic the results can be as different EEG dataset might yield different results for the same framework. Therefore, experimenting with different EEG dataset and testing alternative feature selection schemes can be very interesting for future work.

Machine learning methods for multi-walled carbon nanotubes (MWCNT) genotoxicity prediction.

Multi-walled carbon nanotubes (MWCNTs) are made of multiple single-walled carbon nanotubes (SWCNTs) which are nested inside one another forming concentric cylinders. These nanomaterials are widely used in industrial and biomedical applications, due to their unique physicochemical characteristics. However, previous studies have shown that exposure to MWCNTs may lead to toxicity and some of the physicochemical properties of MWCNTs can influence their toxicological profiles. In silico modelling can be applied as a faster and less costly alternative to experimental (in vivo and in vitro) testing for the hazard characterization of MWCNTs. This study aims at developing a fully validated predictive nanoinformatics model based on statistical and machine learning approaches for the accurate prediction of genotoxicity of different types of MWCNTs. Towards this goal, a number of different computational workflows were designed, combining unsupervised (Principal Component Analysis, PCA) and supervised classification techniques (Support Vectors Machine, “SVM”, Random Forest, “RF”, Logistic Regression, “LR” and Naïve Bayes, “NB”) and Bayesian optimization. The Recursive Feature Elimination (RFE) method was applied for selecting the most important variables. An RF model using only three features was selected as the most efficient for predicting the genotoxicity of MWCNTs, exhibiting 80% accuracy on external validation and high classification probabilities. The most informative features selected by the model were “Length”, “Zeta average” and “Purity”.

Classifying Familial Hypercholesterolaemia: A Tree-based Machine Learning Approach

Familial hypercholesterolaemia is the most common and serious form of inherited hyperlipidaemia. It has an autosomal dominant mode of inheritance, and is characterised by severely elevated low-density lipoprotein cholesterol levels. Familial hypercholesterolaemia is an important cause of premature coronary heart disease, but is potentially treatable. However, the majority of familial hypercholesterolaemia individuals are under-diagnosed and under-treated, resulting in lost opportunities for premature coronary heart disease prevention. This study aims to assess performance of machine learning algorithms for enhancing familial hypercholesterolaemia detection within the Malaysian population. We applied three machine learning algorithms (random forest, gradient boosting and decision tree) to classify familial hypercholesterolaemia among Malaysian patients and to identify relevant features from four well-known diagnostic instruments: Simon Broome, Dutch Lipid Clinic Criteria, US Make Early Diagnosis to Prevent Early Deaths and Japanese FH Management Criteria. The performance of these classifiers was compared using various measurements for accuracy, precision, sensitivity and specificity. Our results indicated that the decision tree classifier had the best performance, with an accuracy of 99.72%, followed by the gradient boosting and random forest classifiers, with accuracies of 99.54% and 99.52%, respectively. The three classifiers with Recursive Feature Elimination method selected six common features of familial hypercholesterolaemia diagnostic criteria (family history of coronary heart disease, low-density lipoprotein cholesterol levels, presence of tendon xanthomata and/or corneal arcus, family hypercholesterolaemia, and family history of familial hypercholesterolaemia) that generate the highest accuracy in predicting familial hypercholesterolaemia. We anticipate machine learning algorithms will enhance rapid diagnosis of familial hypercholesterolaemia by providing the tools to develop a virtual screening test for familial hypercholesterolaemia.

An adaptive ensemble predictive strategy for multiple scale electrical energy usages forecasting

Electrical energy prediction plays an important role in energy management, power plant scheduling, peak demand and grid security conflict. To deal with prediction scenarios at multiple energy and time scales, an adaptive ensemble model strategy with wider applicability is proposed. The strategy consists of three parts. Firstly, the data preprocessing part is composed of expert knowledge, recursive feature elimination (RFE) and fuzzy c-means clustering (FCM). RFE and FCM methods are used for feature identification and data clustering for different energy usage patterns. Secondly, the primary prediction part consists of a model library containing five commonly used data-driven models. To improve their prediction accuracies, key model parameters are optimized by swarm evolution algorithms (EAs). An algorithm package containing three EAs is combined with the model library. The choice of algorithm for each model depends on the comparison of accuracy in specific prediction scenarios. In the third part, a linear regression model provides the final result based on primary predictors’ outputs. Its weights are also optimized by EAs from the algorithm package.To verify the performance of this strategy, four case studies are carried out representing different energy and time scales’ prediction scenarios. Case A is a benchmark case from the first energy prediction competition organized by American Society of Heating Refrigerating and Air-conditioning Engineer (ASHRAE). Case B perform hourly electrical load prediction of a whole building from University of Wyoming, USA. Case C is a city-scale daily electrical load forecasting (Yizheng City, Jiangsu Province, China), and in Case D, national wide monthly electrical load prediction of the United States is carried out. Results of the four cases indicate that: (1) The proposed ensemble strategy performs best in all four case studies compared with five primary predictors. Compared to the best primary predictors in four case studies, the accuracies of the proposed model increase by 3% to 17% (MAPE); (2) Compared with previous reported ensemble models, the accuracy of the proposed model also increase by 11% (MAPE). The adaptive prediction framework is applicable and has the potential to become a general predictive strategy for energy predictions.

Quantitative integration of radiomic and genomic data improves survival prediction of low-grade glioma patients.

Glioma is the most common and most serious form of brain tumors that affects adults. Accurate prediction of survival and phenotyping of low-grade glioma (LGG) patients at high or low risk are the key to achieving precision diagnosis and treatment. This study is aimed to integrate both magnetic resonance imaging (MRI) data and gene expression data to develop a new integrated measure that represents a LGG patient's disease-specific survival (DSS) and classify subsets of patients at low and high risk for progression to cancer. We first construct the gene regulatory network by using gene expression data. We obtain twelve network modules and identify eight image biomarkers by using the Cox regression model with MRI data. Furthermore, correlation analysis between gene modules and image features identify four radiomic features. The least absolute shrinkage and selection operator (Lasso) method is applied to predict these image features with gene expression data when lacking MRI data or image segmentation technology. Furthermore, the support vector machine (SVM)-based recursive feature elimination method has been established to predict DSS using gene signatures. Finally, 4 image signatures and 43 gene signatures are recognized to be associated with the patient's prognosis. An integrated measure for combining image and gene signatures is obtained through the PSO algorithm. The concordance index (C-index) and the time-dependent receiver operating characteristic (ROC) analysis are used to evaluate prediction accuracy. The C-index obtained for this integrated measure is 0.8071 and the area under the curve (AUC) of the ROC curve is 0.79, which are higher than any other single features. The 72.1% accuracy of classification of patients is better than the accuracy associated with the published work. These results demonstrate that integration analysis of radiomic and genomic data can improve the accuracy of the prediction of survival for lower grade gliomas.

Wind Farm Fault Detection by Monitoring Wind Speed in the Wake Region

A novel concept of wind farm fault detection by monitoring the wind speed in the wake region is proposed in this study. A wind energy dissipation model was coupled with a computational fluid dynamics solver to simulate the fluid field of a wind turbine array, and the wind velocity and direction in the simulation were exported for identifying wind turbine faults. The 3D steady Navier–Stokes equations were solved by using the cell center finite volume method with a second order upwind scheme and a k−ε turbulence model. In addition, the wind energy dissipation model, derived from energy balance and Betz’s law, was added to the Navier–Stokes equations’ source term. The simulation results indicate that the wind speed distribution in the wake region contains significant information regarding multiple wind turbine faults. A feature selection algorithm specifically designed for the analysis of wind flow was proposed to reduce the number of features. This algorithm proved to have better performance than fuzzy entropy measures and recursive feature elimination methods under a limited number of features. As a result, faults in the wind turbine array could be detected and identified by machine learning algorithms.

Comparison of Machine Learning Models for Prediction of Initial Intravenous Immunoglobulin Resistance in Children With Kawasaki Disease.

We constructed an optimal machine learning (ML) method for predicting intravenous immunoglobulin (IVIG) resistance in children with Kawasaki disease (KD) using commonly available clinical and laboratory variables. We retrospectively collected 98 clinical records of hospitalized children with KD (2–109 months of age). We found that 20 (20%) children were resistant to initial IVIG therapy. We trained three ML techniques, including logistic regression, linear support vector machine, and eXtreme gradient boosting with 10 variables against IVIG resistance. Moreover, we estimated the predictive performance based on nested 5-fold cross-validation (CV). We also selected variables using the recursive feature elimination method and performed the nested 5-fold CV with selected variables in a similar manner. We compared ML models with the existing system regardless of their predictive performance. Results of the area under the receiver operator characteristic curve were in the range of 0.58–0.60 in the all-variable model and 0.60–0.75 in the select model. The specificities were more than 0.90 and higher than those in existing scoring systems, but the sensitivities were lower. Three ML models based on demographics and routine laboratory variables did not provide reliable performance. This is possibly the first study that has attempted to establish a better predictive model. Additional biomarkers are probably needed to generate an effective prediction model.

Towards a novel set of GPS‐derived metrics to identify the differences between mobility patterns of cognitively intact older adults and older adults with dementia

AbstractBackgroundMaintaining an active lifestyle and participating in social activities are key components of healthy ageing. These components rely on the individuals’ ability to remain mobile out of home. The global positioning system (GPS) is increasingly used to assess outdoor mobility of older adults. However, there is a gap in establishing a framework that recognizes the differences between mobility patterns of cognitively intact older adults and older adults with dementia (OAwD). We propose a set of GPS‐derived metrics to identify these differences, which can be used to evaluate changes in mobility patterns of OAwD over time.MethodWe analyzed the mobility profiles of 15 older adults from the greater Toronto area using GPS technology. Participants were aged between 65 and 90 years (M=74.9, SD=7.05) and were either cognitively healthy (n = 8) or diagnosed with dementia (n = 7). First, we created GPS‐based features representing different dimensions of outdoor mobility. These features include (1) the typical distance covered by individuals, (2) randomness of the travel patterns, (3) the spatial variation of the GPS locations, (4) the number of distinct stops visited per day, (5) the number of daily trips away from home, (6) the score of three levels of outdoor life‐space, and (7) the maximum distance travelled from home. Then, we used Recursive Feature Elimination (RFE) with Random Forest (RF) to select the most important mobility features with respect to cognitive status.ResultA total of 55,580 GPS points were collected by the 15 unique participants over a period of 4 to 8 weeks. The RFE method indicated that features (1), (2), (3), and (5) had the strongest association with cognitive status. Our results suggested that OAwD displayed more predictability (i.e. smaller randomness) in travel patterns compared to the controls (OAwD: M=4.82, SD=0.57 vs. CTL: M=5.62, SD=0.71). Furthermore, the OAwD made fewer daily out‐of‐home trips compared to the controls (OAwD: M=1.06, SD=0.79 vs. CTL: M=1.65, SD=0.45), and covered smaller distances compared to the controls (OAwD: M=12.27, SD=11.73 km vs. CTL: M=15.78, SD=24.31 km).ConclusionThe proposed set of GPS‐based features identified the differences between mobility patterns of cognitively intact older adults and OAwD.

The structural similarity index for IMRT quality assurance: radiomics‐based error classification

The implementation of radiomics and machine learning (ML) techniques on analyzing two-dimensional gamma maps has been demonstrated superior to the conventional gamma analysis for error identification in intensity modulated radiotherapy (IMRT) quality assurance (QA). Recently, the Structural SIMilarity (SSIM) sub-index maps were shown to be able to reveal the error types of the dose distributions. In this study, we aimed to apply radiomics analysis on SSIM sub-index maps and develop ML models to classify delivery errors in patient-specific dynamic IMRT QA. Twenty-one sliding-window IMRT plans of 180 beams for three treatment sites were involved in this study. Four types of machine-related errors of various magnitudes were simulated for each beam at each control point, including the monitor unit (MU) variations, same-directional and opposite-directional shifts of the multileaf collimators (MLCs) and random mispositioning of the MLCs. In the QA process, a total of 1620 portal dose (PD) images were acquired for the beams with and without errors. The predicted PD images of the original beams were set as references. To quantify the agreement between a measured PD image and the corresponding predicted PD image, four difference maps including three SSIM sub-index maps, and one dose difference-derived map were calculated. Then, radiomic features were extracted from the four difference maps of each measured PD image. We tested four typical classifiers including linear discriminant classifier (LDC), two supporting vector machine (SVM) classifiers, and random forest (RF) for this multiclass classification task. A nested cross-validation scheme was used for model evaluations, where the SVM recursive feature elimination method was applied for feature selection. Finally, the performance of the ML model on identifying the error-free and the erroneous cases was compared to that of the conventional gamma analysis. The statistics of the selected features showed that all of the difference maps and the feature categories made balanced contributions to solve this classification task. Best performance was achieved by the Linear-SVM model with average overall classification accuracy of 0.86. Specifically, the average classification accuracies of the shift, opening, and the random errors were around 0.9. Moreover, ~80% of error-free and MU errors were correctly classified. Using gamma analysis, the 3mm/3% criterion was found insensitive to errors (sensitivity was only 0.33). Although the sensitivity to errors with the 2mm/2% criterion increased to 0.79, still 8% worse than that of the ML model. We proposed an ML-based method for machine-related error identification in patient-specific dynamic IMRT QA, where radiomic analysis on SSIM sub-index maps were used for feature extraction. With extensive validation to select the best features and classifiers, high accuracies in error classification were achieved. Compared with the conventional gamma threshold method, this approach has great potential in error identification for the patient-specific IMRT QA process.

An Intrusion Detection Method Based on Decision Tree-Recursive Feature Elimination in Ensemble Learning

With the rapid development of the Internet, various forms of network attack have emerged, so how to detect abnormal behavior effectively and to recognize their attack categories accurately have become an important research subject in the field of cyberspace security. Recently, many hot machine learning-based approaches are applied in the Intrusion Detection System (IDS) to construct a data-driven model. The methods are beneficial to reduce the time and cost of manual detection. However, the real-time network data contain an ocean of redundant terms and noises, and some existing intrusion detection technologies have lower accuracy and inadequate ability of feature extraction. In order to solve the above problems, this paper proposes an intrusion detection method based on the Decision Tree-Recursive Feature Elimination (DT-RFE) feature in ensemble learning. We firstly propose a data processing method by the Decision Tree-Based Recursive Elimination Algorithm to select features and to reduce the feature dimension. This method eliminates the redundant and uncorrelated data from the dataset to achieve better resource utilization and to reduce time complexity. In this paper, we use the Stacking ensemble learning algorithm by combining Decision Tree (DT) with Recursive Feature Elimination (RFE) methods. Finally, a series of comparison experiments by cross-validation on the KDD CUP 99 and NSL-KDD datasets indicate that the DT-RFE and Stacking-based approach can better improve the performance of the IDS, and the accuracy for all kinds of features is higher than 99%, except in the case of U2R accuracy, which is 98%.