Epileptic Seizure Detection Based on New Hybrid Models with Electroencephalogram Signals

Objectives: Epileptic seizures are one of the most common diseases in society and difficult to detect. In this study, a new method was proposed to automatically detect and classify epileptic seizures from EEG (Electroencephalography) signals.Methods: In the proposed method, EEG signals classification five-classes including the cases of eyes open, eyes closed, healthy, from the tumor region, an epileptic seizure, has been carried out by using the support vector machine (SVM) and the normalization methods comprising the z-score, minimum-maximum, and MAD normalizations. To classify the EEG signals, the support vector machine classifiers having different kernel functions, including Linear, Cubic, and Medium Gaussian, have been used. In order to evaluate the performance of the proposed hybrid models, the confusion matrix, ROC curves, and classification accuracy have been used. The used SVM models are Linear SVM, Cubic SVM, and Medium Gaussian SVM.Results: Without the normalizations, the obtained classification accuracies are 76.90%, 82.40%, and 81.70% using Linear SVM, Cubic SVM, and Medium Gaussian SVM, respectively. After applying the z-score normalization to the multi-class EEG signals dataset, the obtained classification accuracies are 77.10%, 82.30%, and 81.70% using Linear SVM, Cubic SVM, and Medium Gaussian SVM, respectively. With the minimum-maximum normalization, the obtained classification accuracies are 77.20%, 82.40%, and 81.50% using Linear SVM, Cubic SVM, and Medium Gaussian SVM, respectively. Moreover, finally, after applying the MAD normalization to the multi-class EEG signals dataset, the obtained classification accuracies are 76.70%, 82.50%, and 81.40% using Linear SVM, Cubic SVM, and Medium Gaussian SVM, respectively.Conclusion: The obtained results have shown that the best hybrid model is the combination of cubic SVM and MAD normalization in the classification of EEG signals classification five-classes.     

GPCR-MPredictor: multi-level prediction of G protein-coupled receptors using genetic ensemble

G protein-coupled receptors (GPCRs) are transmembrane proteins, which transduce signals from extracellular ligands to intracellular G protein. Automatic classification of GPCRs can provide important information for the development of novel drugs in pharmaceutical industry. In this paper, we propose an evolutionary approach, GPCR-MPredictor, which combines individual classifiers for predicting GPCRs. GPCR-MPredictor is a web predictor that can efficiently predict GPCRs at five levels. The first level determines whether a protein sequence is a GPCR or a non-GPCR. If the predicted sequence is a GPCR, then it is further classified into family, subfamily, sub-subfamily, and subtype levels. In this work, our aim is to analyze the discriminative power of different feature extraction and classification strategies in case of GPCRs prediction and then to use an evolutionary ensemble approach for enhanced prediction performance. Features are extracted using amino acid composition, pseudo amino acid composition, and dipeptide composition of protein sequences. Different classification approaches, such as k-nearest neighbor (KNN), support vector machine (SVM), probabilistic neural networks (PNN), J48, Adaboost, and Naives Bayes, have been used to classify GPCRs. The proposed hierarchical GA-based ensemble classifier exploits the prediction results of SVM, KNN, PNN, and J48 at each level. The GA-based ensemble yields an accuracy of 99.75, 92.45, 87.80, 83.57, and 96.17% at the five levels, on the first dataset. We further perform predictions on a dataset consisting of 8,000 GPCRs at the family, subfamily, and sub-subfamily level, and on two other datasets of 365 and 167 GPCRs at the second and fourth levels, respectively. In comparison with the existing methods, the results demonstrate the effectiveness of our proposed GPCR-MPredictor in classifying GPCRs families. It is accessible at .     

A new avenue for classification and prediction of olive cultivars using supervised and unsupervised algorithms

Various methods have been used to identify cultivares of olive trees; herein we used different bioinformatics algorithms to propose new tools to classify 10 cultivares of olive based on RAPD and ISSR genetic markers datasets generated from PCR reactions. Five RAPD markers (OPA0a21, OPD16a, OP01a1, OPD16a1 and OPA0a8) and five ISSR markers (UBC841a4, UBC868a7, UBC841a14, U12BC807a and UBC810a13) selected as the most important markers by all attribute weighting models. K-Medoids unsupervised clustering run on SVM dataset was fully able to cluster each olive cultivar to the right classes. All trees (176) induced by decision tree models generated meaningful trees and UBC841a4 attribute clearly distinguished between foreign and domestic olive cultivars with 100% accuracy. Predictive machine learning algorithms (SVM and Naïve Bayes) were also able to predict the right class of olive cultivares with 100% accuracy. For the first time, our results showed data mining techniques can be effectively used to distinguish between plant cultivares and proposed machine learning based systems in this study can predict new olive cultivars with the best possible accuracy.     

Embracing imperfection: Machine-assisted invertebrate classification in real-world datasets

Despite growing concerns over the health of global invertebrate diversity, terrestrial invertebrate monitoring efforts remain poorly geographically distributed. Machine-assisted classification has been proposed as a potential solution to quickly gather large amounts of data; however, previous studies have often used unrealistic or idealized datasets to train and test their models.In this study, we describe a practical methodology for including machine learning in ecological data acquisition pipelines. Here we train and test machine learning algorithms to classify over 72,000 terrestrial invertebrate specimens from morphometric data and contextual metadata. All vouchered specimens were collected in pitfall traps by the National Ecological Observatory Network (NEON) at 45 locations across the United States from 2016 to 2019. Specimens were photographed, and two separate machine learning paradigms were used to classify them. In the first, we used a convolutional neural network (ResNet-50), and in the second, we extracted morphometric data as feature vectors using ImageJ and used traditional machine learning methods to classify specimens. Issues stemming from inconsistent taxonomic label specificity were resolved by making classifications at the lowest identified taxonomic level (LITL). Taxa with too few specimens to be included in the training dataset were classified by the model using zero-shot classification.When classifying specimens that were known and seen by our models, we reached a maximum accuracy of 72.7% using eXtreme Gradient Boosting (XGBoost) at the LITL. This nearly matched the maximum accuracy achieved by the CNN of 72.8% at the LITL. Models that were trained without contextual metadata underperformed models with contextual metadata. We also classified invertebrate taxa that were unknown to the model using zero-shot classification, reaching a maximum accuracy of 65.5% when using the ResNet-50, compared to 39.4% when using XGBoost.The general methodology outlined here represents a realistic application of machine learning as a tool for ecological studies. We found that more advanced and complex machine learning methods such as convolutional neural networks are not necessarily more accurate than traditional machine learning methods. Hierarchical and LITL classifications allow for flexible taxonomic specificity at the input and output layers. These methods also help address the ‘long tail’ problem of underrepresented taxa missed by machine learning models. Finally, we encourage researchers to consider more than just morphometric data when training their models, as we have shown that the inclusion of contextual metadata can provide significant improvements to accuracy.     

Inclusion of fractal dimension in four machine learning algorithms improves the prediction accuracy of mean weight diameter of soil

The study of soil mean weight diameter (MWD), essential for sustainable soil management, has recently received much attention. As the estimation of MWD is challenging, labor-intensive, and time-consuming, there is a crucial need to develop a predictive estimation method to generate helpful information required for the soil health assessment to save time and cost involved in soil analysis. Pedotransfer functions (PTFs) are used to estimate parameters that are ‘difficult to measure’ and time-consuming with the help of ’easy to measure’ parameters. In the current study, empirical PTFs, i.e., multi-linear regression (MLR), and four machine learning based PTFs, i.e., artificial neural network (ANN), support vector machine (SVM), classification and regression trees (CART), and random forest (RF) were used for mean weight diameter prediction in Karnal district of Haryana, India. A total of 121 soil samples from 0‐15 and 15‐30 cm soil depths were collected from seventeen villages of Nilokheri, Nissing, and Assandh blocks of Karnal district. Soil parameters such as bulk density (BD), fractal dimension (D), soil texture (i.e., sand, silt, and clay), organic carbon (OC), and glomalin content were used as the input variables. Two input combinations, i.e., one with texture data (dataset 1) and the other with fractal dimension data replacing texture (dataset 2), were used, and the complete dataset (121) was divided into training and testing datasets in a 4:1 ratio. The model performance was evaluated by statistical parameters such as mean absolute error (MAE), mean absolute percentage error (MAPE), root mean square error (RMSE), normalized root mean square error (NRMSE), and determination coefficient (R²). The comparison results showed that including the fractal dimension in the input dataset improved the prediction capability of ANN, SVM, and RF. MLR and CART showed lower predictive ability than the other three approaches (i.e., ANN, SVM, and RF). In the training dataset, RMSE (mm) for the SVM model was 8.33% lower with D than with texture as the input, whereas, in the testing dataset, it was 16.67% lower. Because SVM is more flexible and effectively captures non-linear relationships, it performed better than the other models in predicting MWD. As seen in this study, the SVM model with input data D is the best in its class and has a high potential for MWD prediction in the Karnal district of Haryana, India.     

The use of machine learning to detect foraging behaviour in whale sharks: a new tool in conservation

In this study we present the first attempt at modelling the feeding behaviour of whale sharks using a machine learning analytical method. A total of eight sharks were monitored with tri-axial accelerometers and their foraging behaviours were visually observed. Our results highlight that the random forest model is a valid and robust approach to predict the feeding behaviour of the whale shark. In conclusion this novel approach exposes the practicality of this method to serve as a conservation tool and the capability it offers in monitoring potential disturbances of the species.     

Using animal-mounted sensor technology and machine learning to predict time-to-calving in beef and dairy cows

Worldwide, there is a trend towards increased herd sizes, and the animal-to-stockman ratio is increasing within the beef and dairy sectors; thus, the time available to monitoring individual animals is reducing. The behaviour of cows is known to change in the hours prior to parturition, for example, less time ruminating and eating and increased activity level and tail-raise events. These behaviours can be monitored non-invasively using animal-mounted sensors. Thus, behavioural traits are ideal variables for the prediction of calving. This study explored the potential of two sensor technologies for their capabilities in predicting when calf expulsion should be expected. Two trials were conducted at separate locations: (i) beef cows (n = 144) and (ii) dairy cows (n = 110). Two sensors were deployed on each cow: (1) Afimilk Silent Herdsman (SHM) collars monitoring time spent ruminating (RUM), eating (EAT) and the relative activity level (ACT) of the cow, and (2) tail-mounted Axivity accelerometers to detect tail-raise events (TAIL). The exact time the calf was expelled from the cow was determined by viewing closed-circuit television camera footage. Machine learning random forest algorithms were developed to predict when calf expulsion should be expected using single-sensor variables and by integrating multiple-sensor data-streams. The performance of the models was tested using the Matthew’s correlation coefficient (MCC), the area under the curve, and the sensitivity and specificity of predictions. The TAIL model was slightly better at predicting calving within a 5-h window for beef cows (MCC = 0.31) than for dairy cows (MCC = 0.29). The TAIL + RUM + EAT models were equally as good at predicting calving within a 5-h window for beef and dairy cows (MCC = 0.32 for both models). Combining data-streams from SHM and tail sensors did not substantially improve model performance over tail sensors alone; therefore, hour-by-hour algorithms for the prediction of time of calf expulsion were developed using tail sensor data. Optimal classification occurred at 2 h prior to calving for both beef (MCC = 0.29) and dairy cows (MCC = 0.25). This study showed that tail sensors alone are adequate for the prediction of parturition and that the optimal time for prediction is 2 h before expulsion of the calf.     

Feature Extraction and Classification of Heart Murmurs Based on Acoustic Qualities

Background and objectiveHeart murmur characterization is a crucial part of cardiac auscultation for determining the potential etiology and severity of heart diseases. One such helpful murmur characterization is the sonic qualities, which reflect both structural and hemodynamical states of the heart. Therefore, the objective is to develop a machine learning based solution for classifying murmur qualities.MethodsFour medically defined murmur qualities, namely the musical quality, blowing-like quality, coarse quality, and soft quality were examined. Feature was extracted from heart murmurs signals in their time domain, frequency domain, time-frequency domain, and phase space domain. Sequential forward floating selection (SFFS) was implemented along with three classifiers, including k-nearest neighbor (KNN), Naïve-Bayes (NB), and linear support vector machine (SVM).ResultsIt was found that multi-domain features are suited for better classification results and linear SVM was able to achieve a better balance between performance and the size of feature subsets among tested classifiers. Using the derived features, classification accuracies of 86%, 91%, 90%, and 84% were achieved for musical quality, blowing-like quality, coarse quality, and soft quality classifications respectively.ConclusionsThe study demonstrated that it is possible to effectively characterize heart murmur through its diagnostic characteristics instead of drawing direct conclusions, which is helpful for retaining versatility and generality found in the conventional cardiac auscultation.     

Predicting protein–protein interactions from sequence using correlation coefficient and high-quality interaction dataset

Identifying protein–protein interactions (PPIs) is critical for understanding the cellular function of the proteins and the machinery of a proteome. Data of PPIs derived from high-throughput technologies are often incomplete and noisy. Therefore, it is important to develop computational methods and high-quality interaction dataset for predicting PPIs. A sequence-based method is proposed by combining correlation coefficient (CC) transformation and support vector machine (SVM). CC transformation not only adequately considers the neighboring effect of protein sequence but describes the level of CC between two protein sequences. A gold standard positives (interacting) dataset MIPS Core and a gold standard negatives (non-interacting) dataset GO-NEG of yeast Saccharomyces cerevisiae were mined to objectively evaluate the above method and attenuate the bias. The SVM model combined with CC transformation yielded the best performance with a high accuracy of 87.94% using gold standard positives and gold standard negatives datasets. The source code of MATLAB and the datasets are available on request under smgsmg@mail.ustc.edu.cn.     

Mem-PHybrid: hybrid features-based prediction system for classifying membrane protein types

Membrane proteins are a major class of proteins and encoded by approximately 20% to 30% of genes in most organisms. In this work, a two-layer novel membrane protein prediction system, called Mem-PHybrid, is proposed. It is able to first identify the protein query as a membrane or nonmembrane protein. In the second level, it further identifies the type of membrane protein. The proposed Mem-PHybrid prediction system is based on hybrid features, whereby a fusion of both the physicochemical and split amino acid composition-based features is performed. This enables the proposed Mem-PHybrid to exploit the discrimination capabilities of both types of feature extraction strategy. In addition, minimum redundancy and maximum relevance has also been applied to reduce the dimensionality of a feature vector. We employ random forest, evidence-theoretic K-nearest neighbor, and support vector machine (SVM) as classifiers and analyze their performance on two datasets. SVM using hybrid features yields the highest accuracy of 89.6% and 97.3% on dataset1 and 91.5% and 95.5% on dataset2 for jackknife and independent dataset tests, respectively. The enhanced prediction performance of Mem-PHybrid is largely attributed to the exploitation of the discrimination power of the hybrid features and of the learning capability of SVM. Mem-PHybrid is accessible at http://www.111.68.99.218/Mem-PHybrid.     

Prediction of neurotoxins based on their function and source

We have developed a method NTXpred for predicting neurotoxins and classifying them based on their function and origin. The dataset used in this study consists of 582 non-redundant, experimentally annotated neurotoxins obtained from Swiss-Prot. A number of modules have been developed for predicting neurotoxins using residue composition based on feed-forwarded neural network (FNN), recurrent neural network (RNN), support vector machine (SVM) and achieved maximum accuracy of 84.19%, 92.75%, 97.72% respectively. In addition, SVM modules have been developed for classifying neurotoxins based on their source (e.g., eubacteria, cnidarians, molluscs, arthropods have been and chordate) using amino acid composition and dipeptide composition and achieved maximum overall accuracy of 78.94% and 88.07% respectively. The overall accuracy increased to 92.10%, when the evolutionary information obtained from PSI-BLAST was combined with SVM module of source classification. We have also developed SVM modules for classifying neurotoxins based on functions using amino acid, dipeptide composition and achieved overall accuracy of 83.11%, 91.10% respectively. The overall accuracy of function classification improved to 95.11%, when PSI-BLAST output was combined with SVM module. All the modules developed in this study were evaluated using five-fold cross-validation technique. The NTXpred is available at www.imtech.res.in/raghava/ntxpred/ and mirror site at http://bioinformatics.uams.edu/mirror/ntxpred.     

Classification success of six machine learning algorithms in radar ornithology

Radar systems have been increasingly used to monitor birds. To take full advantage of the large datasets provided by radars, researchers have implemented machine learning (ML) techniques that automatically read and attempt to classify targets. Here we used data collected from two locations in Portugal with two marine radar antennas (VSR and HSR) to apply and compare the performance of six ML algorithms that are widely used in the literature: random forests (RF), support vector machine (SVM), artificial neural networks (NN), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA) and decision trees (DT), all trained with several dataset configurations. We found that all algorithms performed well (area under the receiver operating characteristic (AUC) and accuracy > 0.80, P < 0.001) when discriminating birds from non‐biological targets such as vehicles, rain or wind turbines, but greater variance in the performance among algorithms was apparent when separating different bird functional groups or bird species (e.g. herons vs. gulls). In our case study, only RF was able to hold an accuracy > 0.80 for all classification tasks, although SVM and DT also performed well. Further, all algorithms correctly classified 86% and 66% (VSR and HSR) of the target points, and only 2% and 4% of these points were misclassified by all algorithms. Our results suggest that ML algorithms are suitable for classifying radar targets as birds, and thereby separating them from other non‐biological targets. The ability of these algorithms to correctly identify among bird species functional groups was found to be much weaker, but if properly trained and supported by a good ground truthing dataset, targeted to the relevant species groups, some of these algorithms are still able to achieve high accuracies in classification tasks. Such results indicate that ML algorithms are suitable for use in near real‐time monitoring of bird movements, and may help to mitigate collision of birds with, for example, wind turbines or airplanes.     

Predicting novel substrates for enzymes with minimal experimental effort with active learning

Enzymatic substrate promiscuity is more ubiquitous than previously thought, with significant consequences for understanding metabolism and its application to biocatalysis. This realization has given rise to the need for efficient characterization of enzyme promiscuity. Enzyme promiscuity is currently characterized with a limited number of human-selected compounds that may not be representative of the enzyme's versatility. While testing large numbers of compounds may be impractical, computational approaches can exploit existing data to determine the most informative substrates to test next, thereby more thoroughly exploring an enzyme's versatility. To demonstrate this, we used existing studies and tested compounds for four different enzymes, developed support vector machine (SVM) models using these datasets, and selected additional compounds for experiments using an active learning approach. SVMs trained on a chemically diverse set of compounds were discovered to achieve maximum accuracies of ~80% using ~33% fewer compounds than datasets based on all compounds tested in existing studies. Active learning-selected compounds for testing resolved apparent conflicts in the existing training data, while adding diversity to the dataset. The application of these algorithms to wide arrays of metabolic enzymes would result in a library of SVMs that can predict high-probability promiscuous enzymatic reactions and could prove a valuable resource for the design of novel metabolic pathways.     

Robust and simplified machine learning identification of pitfall trap‐collected ground beetles at the continental scale

1. Insect populations are changing rapidly, and monitoring these changes is essential for understanding the causes and consequences of such shifts. However, large‐scale insect identification projects are time‐consuming and expensive when done solely by human identifiers. Machine learning offers a possible solution to help collect insect data quickly and efficiently.
2. Here, we outline a methodology for training classification models to identify pitfall trap‐collected insects from image data and then apply the method to identify ground beetles (Carabidae). All beetles were collected by the National Ecological Observatory Network (NEON), a continental scale ecological monitoring project with sites across the United States. We describe the procedures for image collection, image data extraction, data preparation, and model training, and compare the performance of five machine learning algorithms and two classification methods (hierarchical vs. single‐level) identifying ground beetles from the species to subfamily level. All models were trained using pre‐extracted feature vectors, not raw image data. Our methodology allows for data to be extracted from multiple individuals within the same image thus enhancing time efficiency, utilizes relatively simple models that allow for direct assessment of model performance, and can be performed on relatively small datasets.
3. The best performing algorithm, linear discriminant analysis (LDA), reached an accuracy of 84.6% at the species level when naively identifying species, which was further increased to >95% when classifications were limited by known local species pools. Model performance was negatively correlated with taxonomic specificity, with the LDA model reaching an accuracy of ~99% at the subfamily level. When classifying carabid species not included in the training dataset at higher taxonomic levels species, the models performed significantly better than if classifications were made randomly. We also observed greater performance when classifications were made using the hierarchical classification method compared to the single‐level classification method at higher taxonomic levels.
4. The general methodology outlined here serves as a proof‐of‐concept for classifying pitfall trap‐collected organisms using machine learning algorithms, and the image data extraction methodology may be used for nonmachine learning uses. We propose that integration of machine learning in large‐scale identification pipelines will increase efficiency and lead to a greater flow of insect macroecological data, with the potential to be expanded for use with other noninsect taxa.

     

基于支持向量机方法的蛋白可溶性预测

按照蛋白质序列中残基的相对可溶性,将其分为两类(表面／内部)和三类(表面／中间／内部)进行预测。选择不同窗宽和参数对数据进行训练和预测,以确保得到最好的分类效果,并同其他已有方法进行比较。对同一数据集不同分类阈值的预测结果显示,支持向量机方法对蛋白质可溶性的整体预测效果优于神经网络和信息论的方法。其中,对两类数据的最优分类结果达到79．0％,对三类数据的最优分类结果达到67．5％,表明支持向量机是蛋白质残基可溶性预测的一种有效方法。     

Effects of replacing the unreliable cDNA microarray measurements on the disease classification based on gene expression profiles and functional modules

MOTIVATION: Microarrays datasets frequently contain a large number of missing values (MVs), which need to be estimated and replaced for subsequent data mining. The focus of the paper is to study the effects of different MV treatments for cDNA microarray data on disease classification analysis. RESULTS: By analyzing five datasets, we demonstrate that among three kinds of classifiers evaluated in this study, support vector machine (SVM) classifiers are robust to varied MV imputation methods [e.g. replacing MVs by zero, K nearest-neighbor (KNN) imputation algorithm, local least square imputation and Bayesian principal component analysis], while the classification and regression tree classifiers are sensitive in terms of classification accuracy. The KNNclassifiers built on differentially expressed genes (DEGs) are robust to the varied MV treatments, but the performances of the KNN classifiers based on all measured genes can be significantly deteriorated when imputing MVs for genes with larger missing rate (MR) (e.g. MR > 5%). Generally, while replacing MVs by zero performs relatively poor, the other imputation algorithms have little difference in affecting classification performances of the SVM or KNN classifiers. We further demonstrate the power and feasibility of our recently proposed functional expression profile (FEP) approach as means to handle microarray data with MVs. The FEPs, which are derived from the functional modules that are enriched with sets of DEGs and thus can be consistently identified under varied MV treatments, achieve precise disease classification with better biological interpretation. We conclude that the choice of MV treatments should be determined in context of the later approaches used for disease classification. The suggested exclusion criterion of ignoring the genes with larger MR (e.g. >5%), while justifiable for some classifiers such as KNN classifiers, might not be considered as a general rule for all classifiers.     

Detection and classification of chilli leaf disease using a squeeze-and-excitation-based CNN model

Chilli leaf disease has a destructive effect on the chilli crop yield. Chilli leaf disease can result in a significant decrease in both the quantity and quality of the chilli crop. Early detection, perfect identification and accurately diagnosing the disease will aid in increasing the profit of the cultivator. However, after a comprehensive survey investigation, we discovered that no studies have been previously conducted to compare the classification performance of machine learning and deep learning for the chilli leaf disease classification problem. In this study, five main leaf diseases i.e. down curl of a leaf, Geminivirus, Cercospora leaf spot, yellow leaf disease, and up curl disease were identified, and images were captured using a digital camera and are labelled. These diseases were classified using 12 different pretrained deep learning networks (AlexNet, DarkNet53, DenseNet201, EfficientNetb0, InceptionV3, MobileNetV2, NasNetLarge, ResNet101, ShuffleNet, SqueezeNet, VGG19, and XceptionNet) using chilli leaf data with and without augmentation using deep learning transfer. Performance metrics such as accuracy, recall, precision, F1-score, specificity, and misclassification were calculated for each network. VGG19 had the best accuracy (83.54%) without augmentation, and DarkNet53 had the best result (98.82%) with augmentation among all pretrained deep learning networks in our self-built chilli leaf dataset. The result was enhanced by designing a squeeze-and-excitation-based convolutional neural network (SECNN) model. The model was tested on a chilli leaf dataset with different input sizes and mini-batch sizes. The proposed model produced the best accuracy of 98.63% and 99.12% without and with augmentation, respectively. The SECNN model was also tested on different datasets from the PlantVillage data, including apple, cherry, corn, grape, peach, pepper, potato, strawberry, and tomato leaves, separately and with the chilli dataset. The proposed model achieved an accuracy of 99.28% in classifying 43 different classes of plant leaf datasets.     

Radiomics analysis of EPID measurements for patient positioning error detection in thyroid associated ophthalmopathy radiotherapy

PurposeElectronic portal imaging detector (EPID)-based patient positioning verification is an important component of safe radiotherapy treatment delivery. In computer simulation studies, learning-based approaches have proven to be superior to conventional gamma analysis in the detection of positioning errors. To approximate a clinical scenario, the detectability of positioning errors via EPID measurements was assessed using radiomics analysis for patients with thyroid-associated ophthalmopathy.MethodsTreatment plans of 40 patients with thyroid-associated ophthalmopathy were delivered to a solid anthropomorphic head phantom. To simulate positioning errors, combinations of 0-, 2-, and 4-mm translation errors in the left–right (LR), superior-inferior (SI), and anterior-posterior (AP) directions were introduced to the phantom. The positioning errors-induced dose differences between measured portal dose images were used to predict the magnitude and direction of positioning errors. The detectability of positioning errors was assessed via radiomics analysis of the dose differences. Three classification models—support vector machine (SVM), k-nearest neighbors (KNN), and XGBoost—were used for the detection of positioning errors (positioning errors larger or smaller than 3 mm in an arbitrary direction) and direction classification (positioning errors larger or smaller than 3 mm in a specific direction). The receiver operating characteristic curve and the area under the ROC curve (AUC) were used to evaluate the performance of classification models.ResultsFor the detection of positioning errors, the AUC values of SVM, KNN, and XGBoost models were all above 0.90. For LR, SI, and AP direction classification, the highest AUC values were 0.76, 0.91, and 0.80, respectively.ConclusionsCombined radiomics and machine learning approaches are capable of detecting the magnitude and direction of positioning errors from EPID measurements. This study is a further step toward machine learning-based positioning error detection during treatment delivery with EPID measurements.     

Modified Fuzzy Q Learning Based Classifier for Pneumonia and Tuberculosis

This work proposes reinforcement learning for correctly identifying pneumonia and tuberculosis (TB) using a repository of X ray images. To our knowledge, this is a first attempt at employing reinforcement learning for pneumonia and TB classification. In particular, modified fuzzy Q learning (MFQL) algorithm in conjunction with wavelet based pre-processing has been used to build a classifier for identifying pneumonia and tuberculosis's severity. Proposed classifier is a self-learning one and uses pneumonia dataset (no pneumonia, mild pneumonia and severe pneumonia) and tuberculosis dataset (TB present, TB absent) samples to classify X ray images of subjects. Results indicate that MFQL based approach achieves high accuracy and fares much better over contemporary Support Vector Machine (SVM) and k-Nearest Neighbor (KNN) classifiers. Proposed classifier can be a useful tool for pneumonia and tuberculosis diagnosis in a practical setting.