Mathews Correlation Research Articles

Deep-Learning-Based BCI for Automatic Imagined Speech Recognition Using SPWVD

The electroencephalogram (EEG)-based brain–computer interface (BCI) has potential applications in neuroscience and rehabilitation. It benefits a person with neurological impairment to communicate their thoughts to the external world without using any appendages. Decoding imagined speech had limited success, mainly because neural signals are weak and more variable than overt speech, hence challenging to decode by machine-learning (ML)-based algorithms. In recent years, deep learning (DL) with convolutional neural networks (CNNs) has transformed computer vision and can perform pattern recognition better than the traditional ML-based algorithms. The objective of this article is to design a smoothed pseudo-Wigner–Ville distribution (SPWVD) and CNN-based automatic imagined speech recognition (AISR) system to recognize imagined words. This article uses a publically available 64-channel EEG dataset, collected from 15 healthy subjects for three categories: long words, short words, and vowels. The EEG signals were transformed into time–frequency representation (TFR) using SPWVD, which are used as an input to CNN such that the EEG dataset was identified and classified into binary and multiclass categories. In addition, the CNN model was optimized using a recently developed Keras-tuner library to achieve optimal performance. The performance of the SPWVD-CNN-driven AISR system is evaluated using seven performance evaluation metrics: accuracy (ACC), recall (REC), precision (PREC), Mathew’s correlation coefficient (MCC), Cohen’s kappa ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\kappa $ </tex-math></inline-formula> ), F1-score, and area under the curve (AUC). It is found that the proposed system achieved the maximum classification ACCs of 94.82%, 94.26%, 94.68%, and 84.50% for long words, short–long words, short words, and vowels, respectively. The proposed AISR strengthens the possibility of using imagined speech recognition as a future BCI application.

Feature selection algorithm based on binary BAT algorithm and optimum path forest classifier for breast cancer detection using both echographic and elastographic mode ultrasound images.

Breast cancer is one of the fatal diseases among women. Every year, its incidence and mortality rate increase globally. Mammography and sonography are widely used in breast cancer detection. Because mammography misses many cancers and shows false negatives in the denser tissues, sonography is preferred to give some extra information in addition to that available from mammography. To improve the performance of breast cancer detection by reducing false positives. The local binary pattern (LBP) texture features must be extracted from ultrasound elastographic and echographic images of the same patients and then fused to form a single feature vector. Local Binary Pattern (LBP) texture features of elastographic and echographic images are extracted, and reduced individually through a hybrid feature selection technique based on binary BAT algorithm (BBA) and optimum path forest (OPF) classifier and then fused serially. Finally, the support vector machine classifier is used to classify the final fused feature set. Various relevant performance metrics such as accuracy, sensitivity, specificity, discriminant power, Mathews correlation coefficient (MCC), F1 score, and Kappa were used to analyze the classification results. The use of LBP feature produces 93.2% accuracy, 94.4% sensitivity, 92.3% specificity, 89.5% precision value, 91.88% F1 score, 93.34% balanced classification rate, and Mathews correlation coefficient of 0.861. The performance was compared with gray level co-occurrence matrix (GLCM), gray level difference matrix (GLDM), and LAWs features, and showed that LBP outperformed. Because the specificity is better, this method could be useful for detecting breast cancer with minimum false negatives.

A New Hybrid Feature Selection Sequence for Predicting Breast Cancer Survivability Using Clinical Datasets

This paper proposes a hybrid feature selection sequence complemented with filter and wrapper concepts to improve the accuracy of Machine Learning (ML) based supervised classifiers for classifying the survivability of breast cancer patients into classes, living and deceased using METABRIC and Surveillance, Epidemiology and End Results (SEER) datasets. The ML-based classifiers used in the analysis are: Multiple Logistic Regression, K-Nearest Neighbors, Decision Tree, Random Forest, Support Vector Machine and Multilayer Perceptron. The workflow of the proposed ML algorithm sequence comprises the following stages: data cleaning, data balancing, feature selection via a filter and wrapper sequence, cross validation-based training, testing and performance evaluation. The results obtained are compared in terms of the following classification metrics: Accuracy, Precision, F1 score, True Positive Rate, True Negative Rate, False Positive Rate, False Negative Rate, Area under the Receiver Operating Characteristics curve, Area under the Precision-Recall curve and Mathews Correlation Coefficient. The comparison shows that the proposed feature selection sequence produces better results from all supervised classifiers than all other feature selection sequences considered in the analysis.

A Machine Learning Framework for Early-Stage Detection of Autism Spectrum Disorders

Autism Spectrum Disorder (ASD) is a type of neurodevelopmental disorder that affects the everyday life of affected patients. Though it is considered hard to completely eradicate this disease, disease severity can be mitigated by taking early interventions. In this paper, we propose an effective framework for the evaluation of various Machine Learning (ML) techniques for the early detection of ASD. The proposed framework employs four different Feature Scaling (FS) strategies i.e., Quantile Transformer (QT), Power Transformer (PT), Normalizer, and Max Abs Scaler (MAS). Then, the feature-scaled datasets are classified through eight simple but effective ML algorithms like Ada Boost (AB), Random Forest (RF), Decision Tree (DT), K-Nearest Neighbors (KNN), Gaussian Naïve Bayes (GNB), Logistic Regression (LR), Support Vector Machine (SVM) and Linear Discriminant Analysis (LDA). Our experiments are performed on four standard ASD datasets (Toddlers, Adolescents, Children, and Adults). Comparing the classification outcomes using various statistical evaluation measures (Accuracy, Receiver Operating Characteristic: ROC curve, F1-score, Precision, Recall, Mathews Correlation Coefficient: MCC, Kappa score, and Log loss), the best-performing classification methods, and the best FS techniques for each ASD dataset are identified. After analyzing the experimental outcomes of different classifiers on feature-scaled ASD datasets, it is found that AB predicted ASD with the highest accuracy of 99.25%, and 97.95% for Toddlers and Children, respectively and LDA predicted ASD with the highest accuracy of 97.12% and 99.03% for Adolescents and Adults datasets, respectively. These highest accuracies are achieved while scaling Toddlers and Children with normalizer FS and Adolescents and Adults with the QT FS method. Afterward, the ASD risk factors are calculated, and the most important attributes are ranked according to their importance values using four different Feature Selection Techniques (FSTs) i.e., Info Gain Attribute Evaluator (IGAE), Gain Ratio Attribute Evaluator (GRAE), Relief F Attribute Evaluator (RFAE), and Correlation Attribute Evaluator (CAE). These detailed experimental evaluations indicate that proper finetuning of the ML methods can play an essential role in predicting ASD in people of different ages. We argue that the detailed feature importance analysis in this paper will guide the decision-making of healthcare practitioners while screening ASD cases. The proposed framework has achieved promising results compared to existing approaches for the early detection of ASD.

SCADA intrusion detection scheme exploiting the fusion of modified decision tree and Chi-square feature selection

The industrial internet of things (IIoT) and supervisory control and data acquisition (SCADA) have experienced ubiquitous growth recently. This growth comes with the challenge of an increased number of unusual attacks constituting threats. The existence and effect of intruders and their innovative attack techniques are rising. Although the existing intrusion detection systems (IDS) safeguard the networks, they have been computationally expensive. In real-time domains, available methods lag, necessitating additional research into effective feature extraction schemes with time exigency. An IDS with a fused feature selection (FS) approach for detecting and classifying attacks in a real-time SCADA network is imperative. It is to enable the resolution of computationally complex vulnerability detection schemes. The proposed technique is in three (3) phases: (a) data preparation which involves data cleansing and normalization, and (b) a fused feature selection approach built to obtain an optimal subset of features using Chi-square. (c) deployment of the modified decision tree (MDT) for anomaly detection and classification. Lastly, the reliability of the proposed model was validated, demonstrating suitability in precisely detecting abnormalities while minimizing computational time. This improvement enables adaptability for the IDS deployment scheme in a real-time situation, which could be in the control center. The validation results reveal that when the proposed chi-square-based (fused) feature extraction is employed, it performs optimally to other FS techniques and ML classifiers, compared across four (4) publicly available datasets. Cohen’s kappa coefficient (CKC) further validates the proposed model’s reliability. Further demonstrating the experimental results with recourse to false positive rates (FPR), the Mathews correlation coefficient (MCC) was employed. It also shows the resilience of the proposed model performance on an imbalanced dataset validating its suitability in real scenarios.

Using Recurrent Neural Networks for Predicting Type-2 Diabetes from Genomic and Tabular Data.

The development of genomic technology for smart diagnosis and therapies for various diseases has lately been the most demanding area for computer-aided diagnostic and treatment research. Exponential breakthroughs in artificial intelligence and machine intelligence technologies could pave the way for identifying challenges afflicting the healthcare industry. Genomics is paving the way for predicting future illnesses, including cancer, Alzheimer's disease, and diabetes. Machine learning advancements have expedited the pace of biomedical informatics research and inspired new branches of computational biology. Furthermore, knowing gene relationships has resulted in developing more accurate models that can effectively detect patterns in vast volumes of data, making classification models important in various domains. Recurrent Neural Network models have a memory that allows them to quickly remember knowledge from previous cycles and process genetic data. The present work focuses on type 2 diabetes prediction using gene sequences derived from genomic DNA fragments through automated feature selection and feature extraction procedures for matching gene patterns with training data. The suggested model was tested using tabular data to predict type 2 diabetes based on several parameters. The performance of neural networks incorporating Recurrent Neural Network (RNN) components, Long Short-Term Memory (LSTM), and Gated Recurrent Units (GRU) was tested in this research. The model's efficiency is assessed using the evaluation metrics such as Sensitivity, Specificity, Accuracy, F1-Score, and Mathews Correlation Coefficient (MCC). The suggested technique predicted future illnesses with fair Accuracy. Furthermore, our research showed that the suggested model could be used in real-world scenarios and that input risk variables from an end-user Android application could be kept and evaluated on a secure remote server.

Machine learning‐based prediction of cognitive outcomes in de novo Parkinson's disease

AbstractBackgroundCognitive impairment is a debilitating symptom in Parkinson’s disease (PD), with high variability in onset and course of progression. This present study aimed to establish an accurate multivariate machine learning model to predict cognitive outcome in newly diagnosed PD.MethodUsing baseline measures from the Parkinson’s Progression Markers Initiative (PPMI) cohort we subset clinical, biofluid and genetic/epigenetic variables. Annual cognitive assessments over an eight‐year time span were used to define two outcomes of i) dementia conversion, and ii) cognitive impairment. Outcomes and variable subsets were tested using multiple machine learning (ML) algorithms (random forest, elasticnet, SVM‐linear and cforest) to predict prognosis of individual cognitive decline in PD. Prediction was assessed using multiple measures including Area under the Curve (AUC) and Mathews Correlation Coefficient (MCC).ResultFor both cognitive outcomes, irrespective of the of ML algorithm, models consisting of clinical variables alone performed best, with high specificity and the largest AUC's (0.88 ‐ 0.92) and MCC's (0.57‐0.80). Notably, cognitive impairment outcome showed better sensitivity than dementia conversion outcome (0.72 – 0.81 vs 0.29 – 0.64, respectively). Addition of biological and genetic variables did not largely improve the model performance. However, a number of cerebrospinal fluid (CSF) proteins and epigenetic markers showed high predictive weighting in multiple models, when included alongside clinical variables.ConclusionMachine learning algorithms can accurately predict cognitive impairment development in PD. Within the generated models, clinical predictors appear to play a more prominent role than biological predictors. We further present additional future study data, expanding epigenetic research into PD dementia.

An improved hawks optimizer based learning algorithms for cardiovascular disease prediction

Earlier Cardiovascular Disease (CVD) prediction is difficult and the prediction complexity is higher due to the lack of intelligent models. This research proposes a stacked ensemble model for feature extraction/selection and classification process. A novel sample-based Neural Network Reasoning model is employed for classification purposes and the construction of the stacking model. The model's overall performance is optimized using Hawks Optimizer (HO) to acquire optimal and global solution. The dataset for predicting the CVD is obtained from the Kaggle dataset for cardiovascular disease prediction, an online resource. The class imbalance issues encountered in the dataset are also considered as an essential factor that needs to resolve for enhancing the prediction quality. Here, the simulation is performed using MATLAB 2020a and the experimental outcomes show the significance of the model. Some performance metrics like Mathews Correlation Coefficient (MCC), accuracy, precision, F-measure, etc. The comparison is made with various existing benchmark datasets. The anticipated model provides better trade-off in contrast to the existing approaches.

MP-VHPPI: Meta predictor for viral host protein-protein interaction prediction in multiple hosts and viruses.

Viral-host protein-protein interaction (VHPPI) prediction is essential to decoding molecular mechanisms of viral pathogens and host immunity processes that eventually help to control the propagation of viral diseases and to design optimized therapeutics. Multiple AI-based predictors have been developed to predict diverse VHPPIs across a wide range of viruses and hosts, however, these predictors produce better performance only for specific types of hosts and viruses. The prime objective of this research is to develop a robust meta predictor (MP-VHPPI) capable of more accurately predicting VHPPI across multiple hosts and viruses. The proposed meta predictor makes use of two well-known encoding methods Amphiphilic Pseudo-Amino Acid Composition (APAAC) and Quasi-sequence (QS) Order that capture amino acids sequence order and distributional information to most effectively generate the numerical representation of complete viral-host raw protein sequences. Feature agglomeration method is utilized to transform the original feature space into a more informative feature space. Random forest (RF) and Extra tree (ET) classifiers are trained on optimized feature space of both APAAC and QS order separate encoders and by combining both encodings. Further predictions of both classifiers are utilized to feed the Support Vector Machine (SVM) classifier that makes final predictions. The proposed meta predictor is evaluated over 7 different benchmark datasets, where it outperforms existing VHPPI predictors with an average performance of 3.07, 6.07, 2.95, and 2.85% in terms of accuracy, Mathews correlation coefficient, precision, and sensitivity, respectively. To facilitate the scientific community, the MP-VHPPI web server is available at https://sds_genetic_analysis.opendfki.de/MP-VHPPI/.

Evaluation of Automatic Prediction of Small Horizontal Curve Attributes of Mountain Roads in GIS Environments

Road curve attributes can be determined by using Geographic Information System (GIS) to be used in road vehicle traffic safety and planning studies. This study involves analyzing the GIS-based estimation accuracy in the length, radius and the number of small horizontal road curves on a two-lane rural road and a forest road. The prediction success of horizontal curve attributes was investigated using digitized raw and generalized/simplified road segments. Two different roads were examined, involving 20 test groups and two control groups, using 22 datasets obtained from digitized and surveyed roads based on satellite imagery, GIS estimates, and field measurements. Confusion matrix tables were also used to evaluate the prediction accuracy of horizontal curve geometry. F-score, Mathews Correlation Coefficient, Bookmaker Informedness and Balanced Accuracy were used to investigate the performance of test groups. The Kruskal–Wallis test was used to analyze the statistical relationships between the data. Compared to the Bezier generalization algorithm, the Douglas–Peucker algorithm showed the most accurate horizontal curve predictions at generalization tolerances of 0.8 m and 1 m. The results show that the generalization tolerance level contributes to the prediction accuracy of the number, curve radius, and length of the horizontal curves, which vary with the tolerance value. Thus, this study underlined the importance of calculating generalizations and tolerances following a manual road digitization.

Protein-DNA Binding Residue Prediction via Bagging Strategy and Sequence-Based Cube-Format Feature.

Protein-DNA interactions play an important role in diverse biological processes. Accurately identifying protein-DNA binding residues is a critical but challenging task for protein function annotations and drug design. Although wet-lab experimental methods are the most accurate way to identify protein-DNA binding residues, they are time consuming and labor intensive. There is an urgent need to develop computational methods to rapidly and accurately predict protein-DNA binding residues. In this study, we propose a novel sequence-based method, named PredDBR, for predicting DNA-binding residues. In PredDBR, for each query protein, its position-specific frequency matrix (PSFM), predicted secondary structure (PSS), and predicted probabilities of ligand-binding residues (PPLBR) are first generated as three feature sources. Secondly, for each feature source, the sliding window technique is employed to extract the matrix-format feature of each residue. Then, we design two strategies, i.e., square root (SR) and average (AVE), to separately transform PSFM-based and two predicted feature source-based, i.e., PSS-based and PPLBR-based, matrix-format features of each residue into three corresponding cube-format features. Finally, after serially combining the three cube-format features, the ensemble classifier is generated via applying bagging strategy to multiple base classifiers built by the framework of 2D convolutional neural network. The computational experimental results demonstrate that the proposed PredDBR achieves an average overall accuracy of 93.7% and a Mathew's correlation coefficient of 0.405 on two independent validation datasets and outperforms several state-of-the-art sequenced-based protein-DNA binding residue predictors. The PredDBR web-server is available at https://jun-csbio.github.io/PredDBR/.

TextConvoNet: a convolutional neural network based architecture for text classification.

This paper presents, TextConvoNet, a novel Convolutional Neural Network (CNN) based architecture for binary and multi-class text classification problems. Most of the existing CNN-based models use one-dimensional convolving filters, where each filter specializes in extracting n-grams features of a particular input word embeddings (Sentence Matrix). These features can be termed as intra-sentence n-gram features. To the best of our knowledge, all the existing CNN models for text classification are based on the aforementioned concept. The presented TextConvoNet not only extracts the intra-sentence n-gram features but also captures the inter-sentence n-gram features in input text data. It uses an alternative approach for input matrix representation and applies a two-dimensional multi-scale convolutional operation on the input. We perform an experimental study on five binary and multi-class classification datasets and evaluate the performance of the TextConvoNet for text classification. The results are evaluated using eight performance measures, accuracy, precision, recall, f1-score, specificity, gmean1, gmean2, and Mathews correlation coefficient (MCC). Furthermore, we extensively compared presented TextConvoNet with machine learning, deep learning, and attention-based models. The experimental results evidenced that the presented TextConvoNet outperformed and yielded better performance than the other used models for text classification purposes.

Testing two NIRs instruments to predict chicken breast meat quality and exploiting machine learning approaches to discriminate among genotypes and presence of myopathies

To discriminate among three poultry meat types (hybrid broiler, hybrid broiler affected by breast myopathies, and slow-growing native breed), and to predict the proximate and the amino acid (AA) composition of breast meat, two NIRs (Near-Infrared) instruments operating between 850 and 2500 nm coupled with chemometric algorithms and Machine Learning (ML) approaches, were tested. The Partial Least Square Discriminant Analysis was performed for genotype identification, resulting in a Mathew Correlation Coefficient (MCC) ranging from 0.61 to 1.00, according to the spectra pretreatments and instrument adopted. The Partial Least Square Regression allowed reaching a high cross-validation determination coefficient (R2cv) for crude protein (0.98) and ether extract (0.99), while only three AA (aspartic acid, alanine and methionine) reached R2cv > 0.55. The latter predictions were successfully used to discriminate between genotypes using Factorial Discriminant Analysis, with an MCC ranging from 0.67 to 0.95. Overall, both tested NIRs instruments allowed to determine the chemical composition of fresh and freeze-dried chicken meat. In this sense, a significant improvement of NIRs data interpretability was achieved thanks to the use of ML algorithms, as it was possible to discriminate the chemical composition of meat depending on the genetic group and the presence of breast myopathies.

A misbehavior detection framework for cooperative intelligent transport systems

With changing times, the need for security increases in all fields, whether we talk about cloud networks or vehicular networks. In every place, it has its importance, but in vehicular networks where the lives of human beings are involved, security becomes the topmost priority. Therefore, this article aims to shed light on Misbehavior Detection Framework (MDF) used in the Cooperative Intelligent Transport Systems community. Here, MDF keeps an eye on malicious entities on the roads. It is done by regularly evaluating two main checks: consistency and local plausibility. These checks are done by Intelligent Transport System Stations. All the messages received through Vehicle-to-Everything are scrutinized through this model. After that, all the messages are evaluated by local detection mechanisms to decide the holistic message’s plausibility. This article mainly focuses on the logic behind the proposed Misbehavior Detection Framework providing more security, evaluating various Machine Learning-based models to ensure one best out of all based on quality and computation latency of all models along with the results of various parameters, such as Recall, Precision, F1 Score, Accuracy, Bookmaker Informedness, Markedness, Mathews Correlation Coefficient, Kappa, and achieved the best results.

Malaria Detection using Microscopic Image Analysis: A Convolution Neural Network Based Approach

Malaria is a potentially fatal disease which is caused by Plasmodium parasite. These parasites are transmitted to humans through the bites of female Anopheles mosquitoes which play the role of disease vector. Five types of plasmodium cause malaria named P. Falciparum, P. Vivax, P. Ovale, P. Knowlesi, and P. Malariae, Among the Plasmodium parasites, Falciparum and Vivax are particularly lethal to humans. Therefore, early detection of malaria is mandatory to avoid the loss of human life. Different automatic/semi-automatic malaria detection techniques are available in the literature, which reduces the chance of human errors in the prognosis of malaria. In recent years, deep learning-based methods have proven to be effective for object detection. Therefore, such methods have caught the attention of researchers to use for the detection of malarial parasites in human blood. In this paper, we proposed a Convolutional Neural Network (CNN) model, which detects malarial parasites in microscopic images of human blood samples with high accuracy. The proposed model comprises 15 layers. It has 8 convolution layers with ReLu activation function, 4 max-pooling layers, 1 flattening layer, and 2 fully connected layers. The proposed method has been evaluated using various statistical measures against existing state-of-the-art methods. The quantitative measures show the effectiveness of the proposed model. It has a 97.42% testing accuracy, 97.42% sensitivity, 97.41% specificity, 97.70% precision, 97.42% recall , 97.97% F1-score , 97.41% Area Under Curve (AUC), and 94.82% Mathews correlation coefficient.

Mathematical Ranking of AI Classifiers Using Confusion Matrix and Matthews Correlation Coefficient

This research primarily deals with mathematically ranking the level of accuracy of various Artificial Intelligence (AI) based machine learning classifiers, using Mathews Correlation Coefficient (MCC), leveraging Confusion Matrices. A detailed Literature survey was done to gather the existing knowledge. This knowledge was used as foundational basis to further build the scope of this research project. The classifiers used were Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Convolutional Neural Network (CNN). These classifiers were used for Face recognition of individuals. A total of 33 adult test subjects with 10 images per subject resulting in a total of 330 distinct images were part of this research project. These test subjects were from the Labeled Faces in The Wild publicly available database along with teachers from James B. Conant High School who voluntarily participated in this research. Python based face recognition program wrappers and associated environment was built around pre-existing classifiers and the image data was passed to these wrappers. These wrappers recognized the faces of individuals in the images over 10 trials wherein each trial consisted of 33 distinct images, with varying degree of accuracy and presented that as outputs. These outputs were used to determine the Confusion Matrices which in turn were used to calculate the MCC scores. The MCC scores were plotted, and these results showed that the SVM AI classifier had the highest level of relative accuracy for face recognition, followed by KNN and CNN classifiers.

Novel logic mining incorporating log linear approach

Mining the best logical rule from the data is a challenging task because not all attribute of the dataset will contribute towards the optimal logical representation. Even if the correct attributes were selected, wrong logical connection in the logical formula will lead to suboptimal logical representation of the datasets. These two factors must be carefully considered in creating more robust logic mining method. In this paper, we proposed a novel logic mining by introducing log-linear analysis to select the best attributes which formulate the logical rule that will be embedded into the energy-based ANN named Discrete Hopfield Neural Network (DHNN). In log-linear phase, the test of the association for each attributes will be carried out where the attributes that have a significant level less than α will be selected before proceeding to the logic mining phase. By using DHNN, the selected attributes via log-linear will be learned and retrieved the optimal induced logic with classification ability. The proposed hybrid model has been tested using various real-life datasets and was compared with several established logic mining methods. Based on the findings, several winning points for the proposed model where the proposed model dominates 3 metrics out of 5 in the average rank. The metrics that achieve the highest average rank are Accuracy (1.800), Precision (3.500), and Mathews Correlation Coefficient (2.700). In accordance with the experimental result obtained, the proposed model has achieved optimal performance with a statistically significant p-value. Hence, these findings lead to an advancement of the existing logic mining via the statistical method.

Identifying the tumor location-associated candidate genes in development of new drugs for colorectal cancer using machine-learning-based approach.

Numerous studies have been conducted to elucidate the relation of tumor proximity to cancer prognosis and treatment efficacy in colorectal cancer. However, the molecular pathways and prognoses of left- and right-sided colorectal cancers are different, and this difference has not been fully investigated at the genomic level. In this study, a set of data science approaches, including six feature selection methods and three classification models, were used in predicting tumor location from gene expression profiles. Specificity, sensitivity, accuracy, and Mathew's correlation coefficient (MCC) evaluation metrics were used to evaluate the classification ability. Gene ontology enrichment analysis was applied by the Gene Ontology PANTHER Classification System. For the most significant 50 genes, protein-protein interactions and drug-gene interactions were analyzed using the GeneMANIA, CytoScape, CytoHubba, MCODE, and DGIdb databases. The highest classification accuracy (90%) is achieved with the most significant 200 genes when the ensemble-decision tree classification model is used with the ReliefF feature selection method. Molecular pathways and drug interactions are investigated for the most significant 50 genes. It is concluded that a machine-learning-based approach could be useful to discover the significant genes that may have an important role in the development of new therapies and drugs for colorectal cancer.

Improved mayfly optimization deep stacked sparse auto encoder feature selection scorched gradient descent driven dropout XLM learning framework for software defect prediction

SummarySoftware testing is the process of improving software quality by classifying and removing defects in the software development. Previously, several methods were used for software defect prediction, but any one method did not provide sufficient accuracy. To overcome this issue, an improved may fly optimization with deep stacked sparse auto encoder feature selection scorched gradient descent driven dropout extreme learning machine framework (SDP‐IMFOFS‐GDDDXLMC) is proposed in this article for software defect prediction. Here, an IMFO method is considered for feature selection techniques. In feature selection technique, features are selected, such as PC1, PC4, and MC1 for selecting the probable minimal attribute. Then, GDDDXLMC approach is used to classify buggy and clean software detection. The proposed approach is implemented in MATLAB platform. The performance metrics, such as accuracy, precision, F‐measure, sensitivity, specificity, Mathew correlation coefficient (MCC) is examined to examine the performance of the proposed method. The simulation performance of the proposed SDP‐IMFOFS‐GDDDXLMC method attains higher accuracy 99.75%, 97.85%, 95.13%, 14.89%, 16.34%, 17.89%, and 98.79, higher sensitivity 96.34%, 91.23%, 89.12%, 12.67%, 17.56%, 18.90%, and 87.25% higher specificity 14.89%, 16.89%, 20.67%, 93.67%, 92.37%, 98.47%, and 94.78% compared with the existing methods, like SDP‐MLP‐PSO, SDP‐BPNN‐RBFNN, SDP‐CNN‐RNN‐LSTM, SDP‐KBN‐LIME, SDP‐SLDeep‐LSTM, SDP‐K‐PCA‐ELM, and SDP‐CNN‐PHI forest, respectively.

Automatic Seizure Detection Using Modified CNN Architecture and Activation Layer

An epileptology expert must visually inspect the EEG to identify abnormal neural activity, which is time-consuming and subject to human errors. The capability of convolution neural networks (CNN) to extract visuospatial features and learn from these discriminative features makes them useful for this task. This paper presents seizure classification based on long-term EEGs using CNN. After filtering, the scalogram is plotted using a 1-second window each. A recently published dataset (TUSZ v1.5.2) was used for the performance evaluation of various CNN-based deep neural networks. The best accuracy obtained for GoogLeNet and AlexNet is 95.88%, and 95.79% respectively with 50 epochs and 32 mini-batch sizes by using the SWISH activation function. The proposed hybrid architecture (AG86) for epoch 50 with mini-batch size 32 has shown the best testing results in terms of accuracy (94.98%) as compared to the SqueezeNet (93.19%), GoogLeNet (92.65%), and AlexNet (94.44%). Similar performance was observed using metrics specificity, sensitivity, Mathew correlation coefficient (MCC), and F1 score. A general inference based on evaluation can be drawn as the proposed hybrid architecture (AG86) showed better test results compared to pre-trained CNN models. Moreover, by replacing ReLU with the SWISH activation function, the performance of AlexNet and GoogLeNet improved.