Discriminating bloom-forming cyanobacteria using lab-based hyperspectral imagery and machine learning: Validation with toxic species under environmental ranges

Cyanobacteria are major contributors to algal blooms in inland waters, threatening ecosystem function and water uses, especially when toxin-producing strains dominate. Here, we examine 140 hyperspectral (HS) images of five representatives of the widespread, potentially toxin-producing and bloom-forming genera Microcystis, Planktothrix, Aphanizomenon, Chrysosporum and Dolichospermum, to determine the potential of utilizing visible and near-infrared (VIS/NIR) reflectance for their discrimination. Cultures were grown under various light and nutrient conditions to induce a wide range of pigment and spectral variability, mimicking variations potentially found in natural environments. Importantly, we assumed a simplified scenario where all spectral variability was derived from cyanobacteria. Throughout the cyanobacterial life cycle, multiple HS images were acquired along with extractions of chlorophyll a and phycocyanin. Images were calibrated and average spectra from the region of interest were extracted using k-means algorithm. The spectral data were pre-processed with seven methods for subsequent integration into Random Forest models, whose performances were evaluated with different metrics on the training, validation and testing sets. Successful classification rates close to 90 % were achieved using either the first or second derivative along with spectral smoothing, identifying important wavelengths in both the VIS and NIR. Microcystis and Chrysosporum were the genera achieving the highest accuracy (>95 %), followed by Planktothrix (79 %), and finally Dolichospermum and Aphanizomenon (>50 %). The potential of HS imagery to discriminate among toxic cyanobacteria is discussed in the context of advanced monitoring, aiming to enhance remote sensing capabilities and risk predictions for water bodies affected by cyanobacterial harmful algal blooms.

Automated classification of attention deficit hyperactivity disorder and conduct disorder using entropy features with ECG signals

BACKGROUND

The most prevalent neuropsychiatric disorder among children is attention deficit hyperactivity disorder (ADHD). ADHD presents with a high prevalence of comorbid disorders such as conduct disorder (CD). The lack of definitive confirmatory diagnostic tests for ADHD and CD make diagnosis challenging. The distinction between ADHD, ADHD + CD and CD is important as the course and treatment are different. Electrocardiography (ECG) signals may become altered in behavioral disorders due to brain-heart autonomic interactions. We have developed a software tool to categorize ADHD, ADHD + CD and CD automatically on ECG signals.

METHOD

ECG signals from participants were decomposed using empirical wavelet transform into various modes, from which entropy features were extracted. Robust ten-fold cross-validation with adaptive synthetic sampling (ADASYN) and z-score normalization were performed at each fold. Analysis of variance (ANOVA) technique was employed to determine the variability within the three classes, and obtained the most discriminatory features. Highly significant entropy features were then fed to classifiers.

RESULTS

Our model yielded the best classification results with the bagged tree classifier: 87.19%, 87.71% and 86.29% for accuracy, sensitivity and specificity, respectively.

CONCLUSION

The proposed expert system can potentially assist mental health professionals in the stratification of the three classes, for appropriate intervention using accessible ECG signals.

Migraine detection from EEG signals using tunable Q-factor wavelet transform and ensemble learning techniques

Migraine is one of the major neurovascular diseases that recur, can persist for a long time, cripple or weaken the brain. This study uses electroencephalogram (EEG) signals for the diagnosis of migraine, and a computer-aided diagnosis system is presented to support expert opinion. A tunable Q-factor wavelet transform (TQWT) based method is proposed for the analysis of the oscillatory structure of EEG signals. With TQWT, EEG signals are decomposed into sub bands. Then, the features are statistically calculated from these bands. The success of the obtained features in distinguishing between migraine patients and healthy control subjects was performed using the Kruskal Wallis test. Feature values ​​obtained from each sub band were classified using well-known ensemble learning techniques and their classification performances were tested. Among the evaluated classifiers, the highest classification performance was achieved as 89.6% by using the Rotation Forest algorithm with the features obtained with Sub band 2. These results reveal the potential of the study as a tool that will support expert opinion in the diagnosis of migraine.

Enhancing P300 based character recognition performance using a combination of ensemble classifiers and a fuzzy fusion method

BACKGROUND

P300-based brain-computer interfaces provide communication pathways without the need for muscle activity by recognizing electrical signals from the brain. The P300 speller is one of the most commonly used BCI applications, as it is very simple and reliable, and it is capable of reaching satisfactory communication performance. However, as with other BCIs, it remains a challenge to improve the P300 speller's performance to increase its practical usability.

NEW METHODS

In this study, we propose a novel multi-feature subset fuzzy fusion (MSFF) framework for the P300 speller to recognize the users' spelling intention. This method includes two parts: 1) feature selection by the Lasso algorithm and feature division; 2) the construction of ensemble LDA classifiers and the fuzzy fusion of those classifiers to recognize user intention.

RESULTS

The proposed framework is evaluated in three public datasets and achieves an average accuracy of 100% after 4 epochs for BCI Competition II Dataset IIb, 96% for BCI Competition III dataset II and 98.3% for the BNCI Horizon Dataset. It indicates that the proposed MSFF method can make use of temporal information of signals and helps to enhance classification performance.

COMPARISON WITH EXISTING METHODS

The proposed MSFF method yields better or comparable performance than previously reported machine learning algorithms.

CONCLUSIONS

The proposed MSFF method is able to improve the performance of P300-based BCIs.

Ensemble-based bag of features for automated classification of normal and COVID-19 CXR images

The medical and scientific communities are currently trying to treat infected patients and develop vaccines for preventing a future outbreak. In healthcare, machine learning is proven to be an efficient technology for helping to combat the COVID-19. Hospitals are now overwhelmed with the increased infections of COVID-19 cases and given patients’ confidentiality and rights. It becomes hard to assemble quality medical image datasets in a timely manner. For COVID-19 diagnosis, several traditional computer-aided detection systems based on classification techniques were proposed. The bag-of-features (BoF) model has shown a promising potential in this domain. Thus, this work developed an ensemble-based BoF classification system for the COVID-19 detection. In this model, we proposed ensemble at the classification step of the BoF. The proposed system was evaluated and compared to different classification systems for different number of visual words to evaluate their effect on the classification efficiency. The results proved the superiority of the proposed ensemble-based BoF for the classification of normal and COVID19 chest X-ray (CXR) images compared to other classifiers.

Ensemble-based model selection for imbalanced data to investigate the contributing factors to multiple fatality road crashes in Ghana

The study aims to identify relevant variables to improve the prediction performance of the crash injury severity (CIS) classification model. Unfortunately, the CIS database is invariably characterized by the class imbalance. For instance, the samples of multiple fatal injury (MFI) severity class are typically rare as opposed to other classes. The imbalance phenomenon may introduce a prediction bias in favour of the majority class and affect the quality of the learning algorithm. The paper proposes an ensemble-based variable ranking scheme that incorporates the data resampling. At the data pre-processing level, majority weighted minority oversampling (MWMOTE) is employed to treat the imbalanced training data. Ensemble of classifiers induced from the balanced data is used to evaluate and rank the individual variables according to their importance to the injury severity prediction. The relevant variables selected are then applied to the balanced data to form a training set for the CIS classification modelling. An empirical comparison is conducted through considering the variable ranking by: 1) the learning of single inductive algorithm with imbalanced data where the relevant variables are applied to the imbalanced data to form the training data; 2) the learning of single inductive algorithm with MWMOTE data and the relevant variables identified are applied to the balanced data to form the training data; and 3) the learning of ensembles with imbalanced data where the relevant variables identified are applied to the imbalanced data to form the training data. Bayesian Networks (BNs) classifiers are then developed for each ranking method, where nested subsets of the top ranked variables are adopted. The model predictions are captured in four performance indicators in the comparative study. Based on three-year (2014-2016) crash data in Ghana, the empirical results show that the proposed method is effective to identify the most prolific predictors of the CIS level. Finally, based on the inference results of BNs developed on the best subset, the study offers the most probable explanations to the occurrence of MFI crashes in Ghana.

Prediction of diabetic protein markers based on an ensemble method

Introduction: A diabetic protein marker is a type of protein that is closely related to diabetes. This kind of protein plays an important role in the prevention and diagnosis of diabetes. Therefore, it is necessary to identify an effective method for predicting diabetic protein markers. In this study, we propose using ensemble methods to predict diabetic protein markers. Methodological issues: The ensemble method consists of two aspects. First, we combine a feature extraction method to obtain mixed features. Next, we classify the protein using ensemble classifiers. We use three feature extraction methods in the ensemble method, including composition and physicochemical features (abbreviated as 188D), adaptive skip gram features (abbreviated as 400D) and g-gap (abbreviated as 670D). There are six traditional classifiers in this study: decision tree, Naive Bayes, logistic regression, part, k-nearest neighbor, and kernel logistic regression. The ensemble classifiers are random forest and vote. First, we used feature extraction methods and traditional classifiers to classify protein sequences. Then, we compared the combined feature extraction methods with single methods. Next, we compared ensemble classifiers to traditional classifiers. Finally, we used ensemble classifiers and combined feature extraction methods to predict samples. Results: The results indicated that ensemble methods outperform single methods with respect to either ensemble classifiers or combined feature extraction methods. When the classifier is a random forest and the feature extraction method is 588D (combined 188D and 400D), the performance is best among all methods. The second best ensemble feature extraction method is 1285D (combining the three methods) with random forest. The best single feature extraction method is 188D, and the worst one is g-gap. Conclusion: According to the results, the ensemble method, either the combined feature extraction method or the ensemble classifier, was better than the single method. We anticipate that ensemble methods will be a useful tool for identifying diabetic protein markers in a cost-effective manner.

Integrative blockwise sparse analysis for tissue characterization and classification

The topic of sparse representation of samples in high dimensional spaces has attracted growing interest during the past decade. In this work, we develop sparse representation-based methods for classification of clinical imaging patterns into healthy and diseased states. We propose a spatial block decomposition method to address irregularities of the approximation problem and to build an ensemble of classifiers that we expect to yield more accurate numerical solutions than conventional sparse analyses of the complete spatial domain of the images. We introduce two classification decision strategies based on maximum a posteriori probability (BBMAP), or a log likelihood function (BBLL) and an approach to adjusting the classification decision criteria. To evaluate the performance of the proposed approach we used cross-validation techniques on imaging datasets with disease class labels. We first applied the proposed approach to diagnosis of osteoporosis using bone radiographs. In this problem we assume that changes in trabecular bone connectivity can be captured by intensity patterns. The second application domain is separation of breast lesions into benign and malignant categories in mammograms. The object classes in both of these applications are not linearly separable, and the classification accuracy may depend on the lesion size in the second application. Our results indicate that the proposed integrative sparse analysis addresses the ill-posedness of the approximation problem and produces very good class separation for trabecular bone characterization and for breast lesion characterization. Our approach yields higher classification rates than conventional sparse classification and previously published convolutional neural networks (CNNs) that we fine-tuned for our datasets, or utilized for feature extraction. The BBLL technique also produced higher classification rates than learners using hand-crafted texture features, and the Bag of Keypoints, which is a sophisticated patch-based method. Furthermore, our comparative experiments showed that the BBLL function may yield more accurate classification than BBMAP, because BBLL accounts for possible estimation bias.

Predicting age from the transcriptome of human dermal fibroblasts

Biomarkers of aging can be used to assess the health of individuals and to study aging and age-related diseases. We generate a large dataset of genome-wide RNA-seq profiles of human dermal fibroblasts from 133 people aged 1 to 94 years old to test whether signatures of aging are encoded within the transcriptome. We develop an ensemble machine learning method that predicts age to a median error of 4 years, outperforming previous methods used to predict age. The ensemble was further validated by testing it on ten progeria patients, and our method is the only one that predicts accelerated aging in these patients.

Minimal ensemble based on subset selection using ECG to diagnose categories of CAN

BACKGROUND AND OBJECTIVE

Early diagnosis of cardiac autonomic neuropathy (CAN) is critical for reversing or decreasing its progression and prevent complications. Diagnostic accuracy or precision is one of the core requirements of CAN detection. As the standard Ewing battery tests suffer from a number of shortcomings, research in automating and improving the early detection of CAN has recently received serious attention in identifying additional clinical variables and designing advanced ensembles of classifiers to improve the accuracy or precision of CAN diagnostics. Although large ensembles are commonly proposed for the automated diagnosis of CAN, large ensembles are characterized by slow processing speed and computational complexity. This paper applies ECG features and proposes a new ensemble-based approach for diagnosis of CAN progression.

METHODS

We introduce a Minimal Ensemble Based On Subset Selection (MEBOSS) for the diagnosis of all categories of CAN including early, definite and atypical CAN. MEBOSS is based on a novel multi-tier architecture applying classifier subset selection as well as the training subset selection during several steps of its operation. Our experiments determined the diagnostic accuracy or precision obtained in 5 × 2 cross-validation for various options employed in MEBOSS and other classification systems.

RESULTS

The experiments demonstrate the operation of the MEBOSS procedure invoking the most effective classifiers available in the open source software environment SageMath. The results of our experiments show that for the large DiabHealth database of CAN related parameters MEBOSS outperformed other classification systems available in SageMath and achieved 94% to 97% precision in 5 × 2 cross-validation correctly distinguishing any two CAN categories to a maximum of five categorizations including control, early, definite, severe and atypical CAN.

CONCLUSIONS

These results show that MEBOSS architecture is effective and can be recommended for practical implementations in systems for the diagnosis of CAN progression.

Multiclassifier Systems for Predicting Neurological Outcome of Patients with Severe Trauma and Polytrauma in Intensive Care Units

This paper presents an ensemble based classification proposal for predicting neurological outcome of severely traumatized patients. The study comprises both the whole group of patients and a subgroup containing those patients suffering traumatic brain injury (TBI). Data was gathered from patients hospitalized in the Intensive Care Unit (ICU) of the University Hospital in Salamanca. Predictive models were induced from both epidemiologic and clinical variables taken at the emergency room and along the stay in the ICU. The large number of variables leads to a low accuracy in the classifiers even when feature selection methods are used. In addition, the presence of a much larger number of instances of one of the classes in the subgroup of TBI patients produces a significantly lesser precision for the minority class. Usual ways of dealing with the last problem is to use undersampling and oversampling strategies, which can lead to the loss of valuable data and overfitting problems respectively. Our proposal for dealing with these problems is based in the use of ensemble multiclassifiers as well as in the use of an ensemble playing the role of base classifier in multiclassifiers. The proposed strategy gave the best values of the selected quality measures (accuracy, precision, sensitivity, specificity, F-measure and area under the Receiver Operator Characteristic curve) as well as the closest values of precision for the two classes under study in the case of the classification from imbalanced data.