Deep Feature Selection Research Articles

Prediction of fire danger index using a new machine learning based method to enhance power system resiliency against wildfires

AbstractWildfires, which can cause significant damage to power systems, are mostly inevitable and unpredictable. Fire danger indexes, such as the Forest Fire Danger Index (FFDI) and the Canadian Fire Weather Index (FWI), measure the potential wildfire danger at a given time and location. Thus, by predicting these fire danger indexes in advance, power system operators can obtain valuable insight into the potential wildfire risks and can better be prepared to tackle the wildfires. However, due to dependency on weather conditions, these indexes usually have volatile time series, which make their prediction complex. Taking these facts into account, this paper, unlike previous approaches that predict fire danger indexes based on climatological models, develops a machine learning‐based forecast process to predict these indexes using the relevant weather data and past performance. To do this, first, a volatility analysis approach is presented to analyse the volatility level of the time series data of a fire danger index. Afterwards, an effective machine learning‐based forecast methodology using a new deep feature selection model is proposed to predict fire danger indexes. The developed forecast methodology is tested on the real‐world data of FFDI and FWI and is compared with several popular alternative methods reported in the literature.

Optimization of grid-connected voltage support technology and intelligent control strategies for new energy stations based on deep learning

To explore the optimization method of grid-connected voltage support technology in new energy stations, this study first analyzes and discusses this technology. Second, this study describes the deep learning model architecture and feature selection in detail and determines the framework used for the optimization model proposed here. Lastly, the development of optimization and control strategies is investigated, and the optimized model’s effectiveness is verified through experiments. The results reveal that the optimized model's accuracy, precision, recall, and F1 score are higher than those of the comparison model in the performance comparison experiment, reaching the highest values of 0.890, 0.888, 0.878, and 0.883, respectively. This reflects that the optimized model shows high performance on small datasets, and its performance benefits become more pronounced as the data volume increases. This feature is particularly significant because, in practical applications, power systems often need to process large amounts of data to achieve efficient voltage support. In simulation experiments, the optimized model demonstrates excellent performance in terms of response time, stability, robustness, and energy consumption. Moreover, this model effectively addresses various data challenges and uncertainties encountered in grid-connected voltage support technology for power systems, thereby providing robust support for stable and efficient voltage regulation. In light of the findings, this study offers substantial insights for advancing research in the realms of power systems and new energy technologies. The exploration into the application of deep learning and intelligent control strategies within power systems reveals significant potential for transforming grid optimization practices. This study accentuates how data-driven methodologies can revolutionize energy management, paving the way for smarter and more efficient energy systems. By enhancing both the responsiveness and operational efficiency of power grids, the study contributes to the acceleration of digital transformation within the energy sector, fostering innovation and laying a robust foundation for future advancements in energy informatics.

A review on advancements in feature selection and feature extraction for high-dimensional NGS data analysis.

Recent advancements in biomedical technologies and the proliferation of high-dimensional Next Generation Sequencing (NGS) datasets have led to significant growth in the bulk and density of data. The NGS high-dimensional data, characterized by a large number of genomics, transcriptomics, proteomics, and metagenomics features relative to the number of biological samples, presents significant challenges for reducing feature dimensionality. The high dimensionality of NGS data poses significant challenges for data analysis, including increased computational burden, potential overfitting, and difficulty in interpreting results. Feature selection and feature extraction are two pivotal techniques employed to address these challenges by reducing the dimensionality of the data, thereby enhancing model performance, interpretability, and computational efficiency. Feature selection and feature extraction can be categorized into statistical and machine learning methods. The present study conducts a comprehensive and comparative review of various statistical, machine learning, and deep learning-based feature selection and extraction techniques specifically tailored for NGS and microarray data interpretation of humankind. A thorough literature search was performed to gather information on these techniques, focusing on array-based and NGS data analysis. Various techniques, including deep learning architectures, machine learning algorithms, and statistical methods, have been explored for microarray, bulk RNA-Seq, and single-cell, single-cell RNA-Seq (scRNA-Seq) technology-based datasets surveyed here. The study provides an overview of these techniques, highlighting their applications, advantages, and limitations in the context of high-dimensional NGS data. This review provides better insights for readers to apply feature selection and feature extraction techniques to enhance the performance of predictive models, uncover underlying biological patterns, and gain deeper insights into massive and complex NGS and microarray data.

Explainable Deep Learning-Based Feature Selection and Intrusion Detection Method on the Internet of Things.

With the rapid advancement of the Internet of Things, network security has garnered increasing attention from researchers. Applying deep learning (DL) has significantly enhanced the performance of Network Intrusion Detection Systems (NIDSs). However, due to its complexity and "black box" problem, deploying DL-based NIDS models in practical scenarios poses several challenges, including model interpretability and being lightweight. Feature selection (FS) in DL models plays a crucial role in minimizing model parameters and decreasing computational overheads while enhancing NIDS performance. Hence, selecting effective features remains a pivotal concern for NIDSs. In light of this, this paper proposes an interpretable feature selection method for encrypted traffic intrusion detection based on SHAP and causality principles. This approach utilizes the results of model interpretation for feature selection to reduce feature count while ensuring model reliability. We evaluate and validate our proposed method on two public network traffic datasets, CICIDS2017 and NSL-KDD, employing both a CNN and a random forest (RF). Experimental results demonstrate superior performance achieved by our proposed method.

Human gender estimation from CT images of skull using deep feature selection and feature fusion

This research endeavors to prognosticate gender by harnessing the potential of skull computed tomography (CT) images, given the seminal role of gender identification in the realm of identification. The study encompasses a corpus of CT images of cranial structures derived from 218 male and 203 female subjects, constituting a total cohort of 421 individuals within the age bracket of 25 to 65 years. Employing deep learning, a prominent subset of machine learning algorithms, the study deploys convolutional neural network (CNN) models to excavate profound attributes inherent in the skull CT images. In pursuit of the research objective, the focal methodology involves the exclusive application of deep learning algorithms to image datasets, culminating in an accuracy rate of 96.4%. The gender estimation process exhibits a precision of 96.1% for male individuals and 96.8% for female individuals. The precision performance varies across different selections of feature numbers, namely 100, 300, and 500, alongside 1000 features without feature selection. The respective precision rates for these selections are recorded as 95.0%, 95.5%, 96.2%, and 96.4%. It is notable that gender estimation via visual radiography mitigates the discrepancy in measurements between experts, concurrently yielding an expedited estimation rate. Predicated on the empirical findings of this investigation, it is inferred that the efficacy of the CNN model, the configurational intricacies of the classifier, and the judicious selection of features collectively constitute pivotal determinants in shaping the performance attributes of the proposed methodology.

Towards Precise Water Quality Assessment : A Deep Learning Approach with Feature Selection in Smart Monitoring Systems

As water quality concerns intensify, the imperative for accurate monitoring systems grows. This study pioneers a novel approach to precise water quality assessment by integrating deep learning techniques and feature selection in smart monitoring systems. Utilizing k-Nearest Neighbors (KNN), Convolutional Neural Network (CNN), and Inception V3 for classification, along with Random Forest, AdaBoost, and XGBoost for feature selection, the study presents a detailed examination of their performance on water quality datasets. The results showcase notable improvements in both training and testing accuracies for KNN when coupled with Random Forest and varying numbers of estimators. The combination of CNN and AdaBoost exhibits robust performance, underscoring the impact of feature extraction on training and testing accuracies. Inception V3, when integrated with XGBoost, demonstrates nuanced results, emphasizing the significance of feature extraction in enhancing classification outcomes. Specifically, the performance metrics reveal a fusion model using XGBoost and Inception V3 achieving an accuracy of 65.82%, surpassing individual models like Inception V3 (60.05%). Similarly, the combination of AdaBoost and CNN attains a performance of 65.17%, outperforming individual models such as CNN (64.32%). Additionally, the integration of Artificial Neural Networks (ANN) with Random Forest yields a performance of 69.05%, showcasing improvement over standalone ANN (55.79%). The findings underscore the efficacy of deep learning models, particularly when integrated with appropriate feature selection algorithms, in advancing the precision of water quality assessment in smart monitoring systems. This study contributes valuable insights to the field of environmental monitoring, providing a basis for further exploration of synergies between deep learning and feature selection for enhanced accuracy in water quality assessment. The proposed approach holds promise for addressing the critical challenge of precise water quality monitoring in the face of escalating environmental concerns.

DeepPIG: deep neural network architecture with pairwise connected layers and stochastic gates using knockoff frameworks for feature selection

Selecting relevant feature subsets is essential for machine learning applications. Among the feature selection techniques, the knockoff filter procedure proposes a unique framework that minimizes false discovery rates (FDR). However, employing a deep neural network architecture for a knockoff filter framework requires higher detection power. Using the knockoff filter framework, we present a Deep neural network with PaIrwise connected layers integrated with stochastic Gates (DeepPIG) for the feature selection model. DeepPIG exhibited better detection power in synthetic data than the baseline and recent models such as Deep feature selection using Paired-Input Nonlinear Knockoffs (DeepPINK), Stochastic Gates (STG), and SHapley Additive exPlanations (SHAP) while not violating the preselected FDR level, especially when the signal of the features were weak. The selected features determined by DeepPIG demonstrated superior classification performance compared with the baseline model in real-world data analyses, including the prediction of certain cancer prognosis and classification tasks using microbiome and single-cell datasets. In conclusion, DeepPIG is a robust feature selection approach even when the signals of features are weak. Source code is available at https://github.com/DMCB-GIST/DeepPIG.

MOX-NET: Multi-stage deep hybrid feature fusion and selection framework for monkeypox classification

Background:Monkeypox virus has quickly expanded throughout several nations, raising serious public health concerns. Lack of precautionary measures raises concerns about the possibility of global pandemic. As a result, early detection of monkeypox is essential for efficient diagnosis and treatment before broad community transmission. Artificial intelligence (AI) and machine learning (ML) have recently shown tremendous promise in early monkeypox diagnosis using image-based diagnoses methods. But, challenges such as low contrast lesions, overloaded datasets, high memory complexity, and redundant feature extraction still need to be addressed. Methods:This work proposes a coalesced CAD system for monkeypox classification using a deep learning framework. The proposed approach includes multiple stages: First, the given dataset images are pre-processed by a proposed fusion-based contrast enhancement method. Second, six pre-trained deep learning models i.e., Vision Transformers (ViT), Shifted Windows (Swin) Transformers, ResNet-50, ResNet-101, EfficientNetV2, and ConvNeXt-V2 are modified and transfer learning is used for training. Third, from all the deep learning networks, the deep feature vectors are acquired and integrated through the convolutional sparse image decomposition fusion strategy. Subsequently, an entropy-controlled firefly feature selection approach is applied to identify the most suitable features for classification. Lastly, the multi-class support vector machine is utilized to classify the input data based on the selected features. Results:The proposed algorithm utilized the MSLID dataset and obtained an accuracy of 98.64% which is better than other state-of-the-art works. Conclusions:Experimental outcomes reveals that the proposed monkeypox classification algorithm outperforms well-established existing algorithms in both visual and quantitative evaluations, resulting in improved accuracy.

A Review on Utilizing Data Mining Techniques for Chronic Kidney Disease Detection

This comprehensive study delves into the application of machine learning (ML) and data mining techniques for the prognosis and diagnosis of Chronic Kidney Disease (CKD), a significant global health concern characterized by the gradual loss of kidney function. Through a detailed examination of various predictive models, the research evaluates the efficacy of different ML algorithms and data mining methodologies in classifying and diagnosing CKD. Utilizing datasets from the UCI machine learning repository and other sources, this study explores a range of ML algorithms-including logistic regression, decision trees, support vector machines, random forest, and deep learning networks-alongside feature selection techniques to enhance prediction accuracy and facilitate early diagnosis. Despite facing challenges such as dataset limitations and the need for external validation, the findings reveal remarkable potential in using ML and data mining to improve CKD diagnosis, with some models achieving accuracy rates exceeding 99%. The research underscores the critical role of technology in advancing CKD diagnosis and management, paving the way for more personalized and effective healthcare solutions.

Detection of COVID-19 Anti-Vaccination from Twitter Data Using Deep Learning and Feature Selection Approaches

The COVID-19 pandemic has evolved into a crisis significantly impacting health, the economy, and social life worldwide. During this crisis, anti-vaccination sentiment poses a considerable obstacle to controlling the epidemic and the effectiveness of vaccination campaigns. This study aimed to detect COVID-19 anti-vaccination sentiment from Twitter data using a combination of deep learning and feature selection approaches. The proposed method integrates a deep learning model with feature selection techniques to identify anti-vaccination sentiment by pinpointing important features in text data. Hybrid TF-IDF and N-gram methods were utilized for feature extraction, followed by Chi-square feature selection. The dataset comprises Twitter text data and two labels. The Synthetic Minority Oversampling Technique (SMOTE) was applied to balance the labels. Long Short-Term Memory (LSTM), a deep learning architecture, was employed for the classification process. The experimental results, obtained by leveraging the proposed feature extraction, feature selection, and LSTM methods, achieved the highest accuracy value of 99.23%. These findings demonstrate the proposed methods' success in effectively detecting COVID-19 anti-vaccination sentiment in text data. The study's results can offer valuable insights for developing health policies and public information strategies, presenting a new and powerful tool for detecting anti-vaccine sentiment in planning vaccination campaigns and public health interventions.

Deep learned features selection algorithm: Removal operation of anomaly feature maps (RO-AFM)

Class/Regression Activation Maps (CAMs/RAMs; AMs) are often embedded into Convolutional Neural Networks (CNNs) for checking activated regions on input images at estimation. CNNs sometime generate unreliable AMs, such as activated regions, are inappropriate. Because AM is calculated by stacking many feature maps generated by the final convolutional layer, when there are Anomaly Feature Maps (AFMs), unreliable AMs can be generated. For example, suppose we have a CNN that evaluates the heart. In this case, the feature maps that focus on regions unrelated to the heart (e.g., shoulders and esophagus) are AFMs. Additionally, we have a hypothesis that the estimation accuracy of CNNs is increased by removing AFMs. However, methods for automatically detecting and removing AFMs have not been sufficiently studied in previous research to improve the performance of CNNs. Therefore, we propose a method named “Removal Operation of Anomaly Feature Maps (RO-AFMs)” to automatically detect and remove AFMs. When applying an RO-AFM to the Global Average Pooling (GAP) feature vectors of a CNN, dimensions of the GAP vector are reduced. Therefore, an RO-AFM is regarded as a deep-feature selection algorithm. From the results of adopting an RO-AFM to a Regression CNN (R-CNN) for estimating pulmonary artery wedge pressure, which is one of the measurement score for representing cardiac anomaly state, improved reliability of AM and estimation accuracy were verified. A comparison of RO-AFM and the existing methods, i.e., Lasso and the Feature Selection Layer (FSL), indicated that RO-AFM performed slightly better on the estimation accuracy. The computation time required for RO-AFM to evaluate all features was 1.833 s on average, confirming that RO-AFM is a lightweight process. Therefore, RO-AFM is useful for constructing a medical CNN that emphasizes explainability (e.g., CNNs for estimating the risk of a disease or a test value from chest X-ray or computed tomography images).

Learning active subspaces and discovering important features with Gaussian radial basis functions neural networks

Providing a model that achieves a strong predictive performance and is simultaneously interpretable by humans is one of the most difficult challenges in machine learning research due to the conflicting nature of these two objectives. To address this challenge, we propose a modification of the radial basis function neural network model by equipping its Gaussian kernel with a learnable precision matrix. We show that precious information is contained in the spectrum of the precision matrix that can be extracted once the training of the model is completed. In particular, the eigenvectors explain the directions of maximum sensitivity of the model revealing the active subspace and suggesting potential applications for supervised dimensionality reduction. At the same time, the eigenvectors highlight the relationship in terms of absolute variation between the input and the latent variables, thereby allowing us to extract a ranking of the input variables based on their importance to the prediction task enhancing the model interpretability. We conducted numerical experiments for regression, classification, and feature selection tasks, comparing our model against popular machine learning models, the state-of-the-art deep learning-based embedding feature selection techniques, and a transformer model for tabular data. Our results demonstrate that the proposed model does not only yield an attractive prediction performance compared to the competitors but also provides meaningful and interpretable results that potentially could assist the decision-making process in real-world applications. A PyTorch implementation of the model is available on GitHub at the following link.11https://github.com/dannyzx/Gaussian-RBFNN.

IFC-GNN: Combining interactions of functional connectivity with multimodal graph neural networks for ASD brain disorder analysis

Many studies now indicate that brain disorders are associated with functional connectivity between brain regions, but the impact of interactions among functional connections on diseases remains to be explored. Moreover, existing models for analyzing the temporal and spatial attributes of fMRI data exhibit limitations in their efficacy. Currently, there is an absence of a comprehensive model capable of simultaneously capturing disease-related interactions within functional connectivity, fully leveraging the temporal and spatial characteristics of fMRI, and achieving superior diagnostic accuracy. To address these gaps, we propose a model to classify brain disorders that employs a novel algorithm and multimodal graph convolutional networks to process the temporal and spatial features of fMRI data separately, along with the demographic information. In the proposed model, a deep feature selection algorithm is employed to discover disease-related biomarkers and evaluate their impact on brain disorders. Remarkably, this approach has led to a classification accuracy of 80.66%, which surpasses the performance of graph neural networks by approximately 10%. Notably, our investigations have uncovered novel biomarkers linked to autism, with our findings corroborated by relevant literature. These achievements underscore the model's potential in revealing critical insights pertinent to autism, thereby facilitating groundbreaking experimental research in the field.

An interpretable deep learning multi-dimensional integration framework for exchange rate forecasting based on deep and shallow feature selection and snapshot ensemble technology

Accurate exchange rate forecasting is important for the better realization of international economic transactions and international currency investments. However, it is challenging to forecast exchange rates due to their high volatility, nonlinearity, and noisy characteristics. This paper creatively proposes a novel interpretable deep learning multi-dimensional integration framework for foreign exchange rate forecasting based on shallow and deep feature selection and snapshot ensemble technology. The main factors affecting exchange rates are considered at a deep and shallow level, and the feature selection project is implemented. A cluster-based multidimensional learning paradigm is proposed. The framework clusters data and constructs predictive sub-models and then embeds snapshot ensemble techniques into models to improve the robustness of single models. Finally, the results are nonlinearly integrated by a deep learning model. In addition, three datasets and twelve comparative models were used to demonstrate the performance of the models. The empirical results show that the proposed model has the smallest mean absolute percentage error (0.407859%, 0.47134%, and 0.470167%, respectively).

Driver state recognition with physiological signals: Based on deep feature fusion and feature selection techniques

Driver state constitutes a significant factor affecting traffic safety, and accurate detection of driver state can significantly enhance driving safety. Therefore, the objective of this investigation is to develop a high accuracy driver state recognition model driven by physiological data. Firstly, expert knowledge and deep learning algorithms were used to extract the shallow and deep features of the driver's electroencephalogram (EEG) and electrodermal activity (EDA) data, respectively; secondly, the principal component analysis (PCA) technique was employed to reduce dimensionality of deep features, enabling fusion with the shallow features; thirdly, an improved recursive feature elimination (RFE) algorithm based on Shapley additive explanation (SHAP) value was proposed for the issue of redundant features in the fused features, and a hybrid feature selection algorithm was devised in conjunction with the chi-square test; finally, the driver state was recognized based on machine learning algorithms. The findings demonstrate superior performance of the driver state recognition model built on the fused features compared to models constructed solely with shallow or deep features. Moreover, the application of the hybrid feature selection algorithm further improved the performance of the driver state recognition model, resulting in a satisfactory accuracy (94.88 %) and F1 score (94.86 %). This study highlights the efficacy of fusing deep features and conducting feature selection in enhancing the accuracy of the driver state recognition model.

Infrared thermal images using PCSAN-Net-DBOA: An approach of breast cancer classification.

This manuscript proposes thermal images using PCSAN-Net-DBOA Initially, the input images are engaged from the database for mastology research with infrared image (DMR-IR) dataset for breast cancer classification. The adaptive distorted Gaussian matched-filter (ADGMF) was used in removing noise and increasing the quality of infrared thermal images. Next, these preprocessed images are given into one-dimensional quantum integer wavelet S-transform (OQIWST) for extracting Grayscale statistic features like standard deviation, mean, variance, entropy, kurtosis, and skewness. The extracted features are given into the pyramidal convolution shuffle attention neural network (PCSANN) for categorization. In general, PCSANN does not show any adaption optimization techniques to determine the optimal parameter to offer precise breast cancer categorization. This research proposes the dung beetle optimization algorithm (DBOA) to optimize the PCSANN classifier that accurately diagnoses breast cancer. The BCD-PCSANN-DBO method is implemented using Python. To classify breast cancer, performance metrics including accuracy, precision, recall, F1 score, error rate, RoC, and computational time are considered. Performance of the BCD-PCSANN-DBO approach attains 29.87%, 28.95%, and 27.92% lower computation time and 13.29%, 14.35%, and 20.54% greater RoC compared with existing methods like breast cancer diagnosis utilizing thermal infrared imaging and machine learning approaches(BCD-CNN), breast cancer classification from thermal images utilizing Grunwald-Letnikov assisted dragonfly algorithm-based deep feature selection (BCD-VGG16) and Breast cancer detection in thermograms using deep selection based on genetic algorithm and Gray Wolf Optimizer (BCD-SqueezeNet), respectively. RESEARCH HIGHLIGHTS: The input images are engaged from the breast cancer dataset for breast cancer classification. The ADQMF was used in removing noise and increasing the quality of infrared thermal images. The extracted features are given into the PCSANN for categorization. DBOA is proposed to optimize PCSANN classifier that classifies breast cancer precisely. The proposed BCD-PCSANN-DBO method is implemented using Python.

Spatio-Temporal Deep Learning-Based Forecasting of Surface Solar Irradiance: Leveraging Satellite Data and Feature Selection

Spatio-Temporal Deep Learning-Based Forecasting of Surface Solar Irradiance: Leveraging Satellite Data and Feature Selection

Deep Curious Feature Selection: A Recurrent, Intrinsic-Reward Reinforcement Learning Approach to Feature Selection

Feature selection (FS) is an important step in the process of building machine learning-based models. The goal of the feature selection step is to find a small subset of features that will provide good prediction results by removing noisy, irrelevant, or repetitive features. Commonly used Wrapper methods use a machine learning model as a black box and its performance as the goal function for evaluating different features' subsets and selecting the best one. To avoid examining all possible subsets (an NP-hard problem), search algorithms are used to find the subsets to be examined, in a heuristic way. As exhaustive search methods are computationally complicated, most methods use simple and greedy search methods that yield only locally optimal results and are not sensitive to possible features interactions, which means that a feature may be chosen at the expense of two others that are more informative together. We analyze the problem of searching the features subset space with reference to two dimensions, namely, memory of past selected features subset and future selected features. We propose a new wrapper feature selection method based on the deep artificial curiosity framework, which implements intrinsic reward reinforcement learning, with long short-term memory unit (LSTM). This novel algorithm integrates these two elements of memory and future step. We show that our method, called the Deep Curious Feature Selection (DCFS) algorithm, deals with feature interactions, and provides a feature subset which improves the accuracy of learning models on artificial and real datasets.

G-ACP: a machine learning approach to the prediction of therapeutic peptides for gastric cancer

Conventional Gastrointestinal (GI) cancer treatments are quite expensive and have major hazards. Nowadays, a different strategy places more emphasis on creating tiny biologically active peptides that do not cause severe poisoning. Anticancer peptides (ACPs) are found through experimental screening, which is time-dependent and frequently fraught with difficulties. Gastric ACPs are emerging as a promising GI cancer treatment in the current day. It is crucial to identify novel gastric ACPs to have an improved knowledge of their functioning processes and treatment of gastric cancer. As a result of the post-genomic era’s massive production of peptide sequences, rapid and effective ACPs using a computational method are essential. Several adaptive statistical techniques for distinguishing ACPs and non-ACPs have recently been developed. A variety of adapted statistically significant methods have been developed to differentiate between ACPs and non-ACPs. Despite significant progress, there is no specific model for the prediction of gastric ACPs because the specific model will predict a particular type of peptide more accurately and quickly. To overcome this, an initiative is taken for the creation of a reliable framework for the accurate identification of gastric ACPs. The current technique in particular contains four possible features along with one hybrid feature encoding mechanisms which are the target-class motif previously indicated by Amino Acid Composition, Dipeptide Composition, Tripeptide Composition (TPC), Pseudo Amino Acid Composition (PAAC), and their Hybrid. Machine Learning algorithms make high-performance and accurate prediction tools. Moreover, highly variable and ideal deep feature selection is done using an ANOVA-based F score for feature pruning. Experiments on a range of algorithms are carried out to identify the optimal operating strategy due to the diverse nature of learning. Following analysis of the empirical results, Naïve Bayes with TPC and Hybrid feature space outperforms other methods with 0.99 accuracy score on the testing dataset. To find the model generalization an external validation is carried out. In external datasets, the Extra Trees with PAAC features outperforms with the accuracy of 0.94. The comparison study shows that our suggested model will predict gastric ACPs more accurately and will be useful in drug development and gastric cancer. The predictive model can be freely accessed at https://github.com/humeraazad10/G-ACP.git. Communicated by Ramaswamy H. Sarma

Application of deep learning and feature selection technique on external root resorption identification on CBCT images.

Artificial intelligence has been proven to improve the identification of various maxillofacial lesions. The aim of the current study is two-fold: to assess the performance of four deep learning models (DLM) in external root resorption (ERR) identification and to assess the effect of combining feature selection technique (FST) with DLM on their ability in ERR identification. External root resorption was simulated on 88 extracted premolar teeth using tungsten bur in different depths (0.5mm, 1mm, and 2mm). All teeth were scanned using a Cone beam CT (Carestream Dental, Atlanta, GA). Afterward, a training (70%), validation (10%), and test (20%) dataset were established. The performance of four DLMs including Random Forest (RF) + Visual Geometry Group 16 (VGG), RF + EfficienNetB4 (EFNET), Support Vector Machine (SVM) + VGG, and SVM + EFNET) and four hybrid models (DLM + FST: (i) FS + RF + VGG, (ii) FS + RF + EFNET, (iii) FS + SVM + VGG and (iv) FS + SVM + EFNET) was compared. Five performance parameters were assessed: classification accuracy, F1-score, precision, specificity, and error rate. FST algorithms (Boruta and Recursive Feature Selection) were combined with the DLMs to assess their performance. RF + VGG exhibited the highest performance in identifying ERR, followed by the other tested models. Similarly, FST combined with RF + VGG outperformed other models with classification accuracy, F1-score, precision, and specificity of 81.9%, weighted accuracy of 83%, and area under the curve (AUC) of 96%. Kruskal Wallis test revealed a significant difference (p = 0.008) in the prediction accuracy among the eight DLMs. In general, all DLMs have similar performance on ERR identification. However, the performance can be improved by combining FST with DLMs.