Classification Model Research Articles

Survey of federated learning in intrusion detection

Intrusion detection methods are crucial means to mitigate network security issues. However, the challenges posed by large-scale complex network environments include local information islands, regional privacy leaks, communication burdens, difficulties in handling heterogeneous data, and storage resource bottlenecks. Federated learning has the potential to address these challenges by leveraging widely distributed and heterogeneous data, achieving load balancing of storage and computing resources across multiple nodes, and reducing the risks of privacy leaks and bandwidth resource demands. This paper reviews the process of constructing federated learning based intrusion detection system from the perspective of intrusion detection. Specifically, it outlines six main aspects: application scenario analysis, federated learning methods, privacy and security protection, selection of classification models, data sources and client data distribution, and evaluation metrics, establishing them as key research content. Subsequently, six research topics are extracted based on these aspects. These topics include expanding application scenarios, enhancing aggregation algorithm, enhancing security, enhancing classification models, personalizing model and utilizing unlabeled data. Furthermore, the paper delves into research content related to each of these topics through in-depth investigation and analysis. Finally, the paper discusses the current challenges faced by research, and suggests promising directions for future exploration.

Automated post-run analysis of arrayed quantitative PCR amplification curves using machine learning

Background The TaqMan Array Card (TAC) is an arrayed, high-throughput qPCR platform that can simultaneously detect multiple targets in a single reaction. However, the manual post-run analysis of TAC data is time consuming and subject to interpretation. We sought to automate the post-run analysis of TAC data using machine learning models. Methods We used 165,214 qPCR amplification curves from two studies to train and test two eXtreme Gradient Boosting (XGBoost) models. Previous manual analyses of the amplification curves by experts in qPCR analysis were used as the gold standard. First, a classification model predicted whether amplification occurred or not, and if so, a second model predicted the cycle threshold (Ct) value. We used 5-fold cross-validation to tune the models and assessed performance using accuracy, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and mean absolute error (MAE). For external validation, we used 1,472 reactions previously analyzed by 17 laboratory scientists as part of an external quality assessment for a multisite study. Results In internal validation, the classification model achieved an accuracy of 0.996, sensitivity of 0.997, specificity of 0.993, PPV of 0.998, and NPV of 0.991. The Ct prediction model achieved a MAE of 0.590. In external validation, the automated analysis achieved an accuracy of 0.997 and a MAE of 0.611, and the automated analysis was more accurate than manual analyses by 14 of the 17 laboratory scientists. Conclusions We automated the post-run analysis of highly-arrayed qPCR data using machine learning models with high accuracy in comparison to a manual gold standard. This approach has the potential to save time and improve reproducibility in laboratories using the TAC platform and other high-throughput qPCR approaches.

Testing of rapid evaporative mass spectrometry for histological tissue classification and molecular diagnostics in a multi-site study.

While REIMS technology has successfully been demonstrated for the histological identification of ex-vivo breast tumor tissues, questions regarding the robustness of the approach and the possibility of tumor molecular diagnostics still remain unanswered. In the current study, we set out to determine whether it is possible to acquire cross-comparable REIMS datasets at multiple sites for the identification of breast tumors and subtypes. A consortium of four sites with three of them having access to fresh surgical tissue samples performed tissue analysis using identical REIMS setups and protocols. Overall, 21 breast cancer specimens containing pathology-validated tumor and adipose tissues were analyzed and results were compared using uni- and multivariate statistics on normal, WT and PIK3CA mutant ductal carcinomas. Statistical analysis of data from standards showed significant differences between sites and individual users. However, the multivariate classification models created from breast cancer data elicited 97.1% and 98.6% correct classification for leave-one-site-out and leave-one-patient-out cross validation. Molecular subtypes represented by PIK3CA mutation gave consistent results across sites. The results clearly demonstrate the feasibility of creating and using global classification models for a REIMS-based margin assessment tool, supporting the clinical translatability of the approach.

Cross-region feature fusion of global and local area for subtype classification prediction in cervical tumour

Background and objectivesTo develop a cross-regional feature fusion model for the classification of cervical tumour subtypes using non-invasive magnetic resonance imaging (MRI), we aimed to explore feature representation in both global and local areas, and to compare their effect on predictive performance. Method and materialsThis retrospective study included 100 patients with cervical cancer, approved by the Ethical Review Board Committee. Self-supervised learning-based global features were fused with local features for subtype classification modelling. Global features were extracted from the bottleneck of our 3D autoencoder network, while local features were derived based on a radiomics tool. Utilizing the global and local features, the classification model is based on machine learning algorithms to predict two subtypes with pathologically confirmed cervical squamous cell carcinoma and adenocarcinoma. Comparison performance was accessed using area under the curve (AUC), sensitivity, specificity, F1 score, and precision. ResultThe cross-regional feature fusion model showed the best performance (accuracy: 0.95 vs 0.65 in the fusion model and global model) by the support vector machines (SVM) classifier. Even when applied to axial slices with various classification methods, the fusion model consistently yields the best results. ConclusionOur approach preliminary evidence suggests that the fusion of global and local features provides a significant advantage in the clinical diagnosis of cervical cancer subtypes, warranting further investigation and potential application in cervical cancer diagnosis.

Improving the Accuracy and Effectiveness of Text Classification Based on the Integration of the Bert Model and a Recurrent Neural Network (RNN_Bert_Based)

This paper proposes a new robust model for text classification on the Stanford Sentiment Treebank v2 (SST-2) dataset in terms of model accuracy. We developed a Recurrent Neural Network Bert based (RNN_Bert_based) model designed to improve classification accuracy on the SST-2 dataset. This dataset consists of movie review sentences, each labeled with either positive or negative sentiment, making it a binary classification task. Recurrent Neural Networks (RNNs) are effective for text classification because they capture the sequential nature of language, which is crucial for understanding context and meaning. Bert excels in text classification by providing bidirectional context, generating contextual embeddings, and leveraging pre-training on large corpora. This allows Bert to capture nuanced meanings and relationships within the text effectively. Combining Bert with RNNs can be highly effective for text classification. Bert’s bidirectional context and rich embeddings provide a deep understanding of the text, while RNNs capture sequential patterns and long-range dependencies. Together, they leverage the strengths of both architectures, leading to improved performance on complex classification tasks. Next, we also developed an integration of the Bert model and a K-Nearest Neighbor based (KNN_Bert_based) method as a comparative scheme for our proposed work. Based on the results of experimentation, our proposed model outperforms traditional text classification models as well as existing models in terms of accuracy.

Simple and accurate genomic classification model for distinguishing between human and pig Staphylococcus aureus

Staphylococcus aureus (S. aureus) can cause various infections in humans and animals, contributing to high morbidity and mortality. To prevent and control cross-species transmission of S. aureus, it is necessary to understand the host-associated genetic variants. We performed a two-stage genome-wide association study (GWAS) including initial screening and further validation to compare genomic differences between human and pig S. aureus, aiming to identify host-associated determinants. Our multiple GWAS analyses found six consensus significant k-mers associated with host species, providing novel genetic evidence for distinguishing human from pig S. aureus. The best k-mer predictor achieved a high classification accuracy of 98.12% on its own and had extremely high resolution similar to the SNPs-based phylogeny, offering a very simple target for predicting the cross-species transmission risk of S. aureus. The final k-mer model revealed that 90% of S. aureus isolates from farm workers were predicted as livestock origin, suggesting a high risk of cross-species transmission. Bayesian inference revealed different cross-species transmission directions, with the human-to-pig transmission for ST5 and the pig-to-human transmission for ST398. Our findings provide a simple and accurate k-mer model for identifying and predicting the cross-species transmission risk of S. aureus.

Investigation of protein family relationships with deep learning

Abstract Motivation In this paper, we propose a method for finding similarity between Pfam families based on the pre-trained neural network ProtENN2. We use the model ProtENN2 per-residue embeddings to produce new high-dimensional per-family embeddings and develop an approach for calculating inter-family similarity scores based on these embeddings, and evaluate its predictions using structure comparison. Results We apply our method to Pfam annotation by refining clan membership for Pfam families, suggesting both new members of existing clans and potential new clans for future Pfam releases. We investigate some of the failure modes of our approach, which suggests directions for future improvements. Our method is relatively simple with few parameters and could be applied to other protein family classification models. Overall, our work suggests potential benefits of employing deep learning for improving our understanding of protein family relationships and functions of previously uncharacterized families. Availability Github.com/iponamareva/ProtCNNSim, 10.5281/zenodo.10091909 Supplementary information Supplementary data are available at Bioinformatics Advances online.

Investigating the contribution of image time series observations to cauliflower harvest-readiness prediction

Cauliflower cultivation is subject to high-quality control criteria during sales, which underlines the importance of accurate harvest timing. Using time series data for plant phenotyping can provide insights into the dynamic development of cauliflower and allow more accurate predictions of when the crop is ready for harvest than single-time observations. However, data acquisition on a daily or weekly basis is resource-intensive, making selection of acquisition days highly important. We investigate which data acquisition days and development stages positively affect the model accuracy to get insights into prediction-relevant observation days and aid future data acquisition planning. We analyze harvest-readiness using the cauliflower image time series of the GrowliFlower dataset. We use an adjusted ResNet18 classification model, including positional encoding of the data acquisition dates to add implicit information about development. The explainable machine learning approach GroupSHAP analyzes time points' contributions. Time points with the lowest mean absolute contribution are excluded from the time series to determine their effect on model accuracy. Using image time series rather than single time points, we achieve an increase in accuracy of 4%. GroupSHAP allows the selection of time points that positively affect the model accuracy. By using seven selected time points instead of all 11 ones, the accuracy improves by an additional 4%, resulting in an overall accuracy of 89.3%. The selection of time points may therefore lead to a reduction in data collection in the future.

Content-Based Image Retrieval and Feature Extraction: Analysing the Literature

A significant amount of multimedia data consists of digital images, and multimedia content analysis is used in many real-world computer vision applications. Multimedia information, especially photos, has become much more complicated in the last several years. Every day, millions of photos are posted to various websites, such as Instagram, Facebook, and Twitter. Finding a suitable image in an archive is a difficult research subject for the field of computer vision. Most search engines use standard text-based techniques that depend on metadata and captions in order to fetch photos. Over the past 20 years, a great deal of research has been conducted on content-based image retrieval (CBIR), picture categorization, and analysis. In image classification models and CBIR, high-level picture representations are represented as feature vectors made up of numerical values. Empirical evidence indicates a considerable disparity between picture feature representation and human visual understanding. Reducing the semantic gap between human visual understanding and picture feature representation is the aim of this study. This study aims to do a thorough analysis of the latest advancements in the domains of Content-Based picture Retrieval and picture representation. We performed a comprehensive analysis of many models for image retrieval and picture representation, encompassing the most recent advancements in semantic deep-learning methods and feature extraction. This paper provides an in-depth analysis of the key ideas and important studies related to image representation and content-based picture retrieval. In an effort to stimulate more research in this field, it also offers a preview of potential future study topics.

Enhanced classification and severity prediction of major depressive disorder using acoustic features and machine learning

BackgroundPrevious studies have classified major depression and healthy control groups based on vocal acoustic features, but the classification accuracy needs to be improved. Therefore, this study utilized deep learning methods to construct classification and prediction models for major depression and healthy control groups.Methods120 participants aged 16–25 participated in this study, included 64 MDD group and 56 HC group. We used the Covarep open-source algorithm to extract a total of 1200 high-level statistical functions for each sample. In addition, we used Python for correlation analysis, and neural network to establish the model to distinguish whether participants experienced depression, predict the total depression score, and evaluate the effectiveness of the classification and prediction model.ResultsThe classification modelling of the major depression and the healthy control groups by relevant and significant vocal acoustic features was 0.90, and the Receiver Operating Characteristic (ROC) curves analysis results showed that the classification accuracy was 84.16%, the sensitivity was 95.38%, and the specificity was 70.9%. The depression prediction model of speech characteristics showed that the predicted score was closely related to the total score of 17 items of the Hamilton Depression Scale(HAMD-17) (r=0.687, P<0.01); and the Mean Absolute Error(MAE) between the model’s predicted score and total HAMD-17 score was 4.51.LimitationThis study’s results may have been influenced by anxiety comorbidities.ConclusionThe vocal acoustic features can not only effectively classify the major depression and the healthy control groups, but also accurately predict the severity of depressive symptoms.

Hyperspectral Imaging for Phenotyping Plant Drought Stress and Nitrogen Interactions Using Multivariate Modeling and Machine Learning Techniques in Wheat

Accurate detection of drought stress in plants is essential for water use efficiency and agricultural output. Hyperspectral imaging (HSI) provides a non-invasive method in plant phenotyping, allowing the long-term monitoring of plant health due to sensitivity to subtle changes in leaf constituents. The broad spectral range of HSI enables the development of different vegetation indices (VIs) to analyze plant trait responses to multiple stresses, such as the combination of nutrient and drought stresses. However, known VIs may underperform when subjected to multiple stresses. This study presents new VIs in tandem with machine learning models to identify drought stress in wheat plants under varying nitrogen (N) levels. A pot wheat experiment was set up in the glasshouse with four treatments: well-watered high-N (WWHN), well-watered low-N (WWLN), drought-stress high-N (DSHN) and drought-stress low-N (DSLN). In addition to ensuring that plants were watered according to the experiment design, photosynthetic rate (Pn) and stomatal conductance (gs) (which are used to assess plant drought stress) were taken regularly, serving as the ground truth data for this study. The proposed VIs, together with known VIs, were used to train three classification models: support vector machines (SVM), random forest (RF), and deep neural networks (DNN) to classify plants based on their drought status. The proposed VIs achieved more than 0.94 accuracy across all models, and their performance further increased when combined with known VIs. The combined VIs were used to train three regression models to predict the stomatal conductance and photosynthetic rates of plants. The random forest regression model performed best, suggesting that it could be used as a stand-alone tool to forecast gs and Pn and track drought stress in wheat. This study shows that combining hyperspectral data with machine learning can effectively monitor and predict drought stress in crops, especially in varying nitrogen conditions.

Machine-Learning Predictions of Critical Temperatures from Chemical Compositions of Superconductors.

In the quest for advanced superconducting materials, the accurate prediction of critical temperatures (Tc) poses a formidable challenge, largely due to the complex interdependencies between superconducting properties and the chemical and structural characteristics of a given material. To address this challenges, we have developed a machine-learning framework that aims to elucidate these complicated and hitherto poorly understood structure-property and property-property relationships. This study introduces a novel machine-learning-based workflow, termed the Gradient Boosted Feature Selection (GBFS), which has been tailored to predict Tc for superconductors by employing a distributed gradient-boosting framework. This approach integrates exploratory data analyses, statistical evaluations, and multicollinearity reduction techniques to select highly relevant features from a high-dimensional feature space, derived solely from the chemical composition of materials. Our methodology was rigorously tested on a data set comprising approximately 16,400 chemical compounds with around 12,000 unique chemical compositions. The GBFS workflow enabled the development of a classification model that distinguishes compositions likely to exhibit Tc values greater than 10 K. This model achieved a weighted average F1-score of 0.912, an AUC-ROC of 0.986, and an average precision score of 0.919. Additionally, the GBFS workflow underpinned a regression model that predicted Tc values with an R2 of 0.945, an MAE of 3.54 K, and an RMSE of 6.57 K on a test set obtained via random splitting. Further exploration was conducted through out-of-sample Tc predictions, particularly those exceeding the liquid nitrogen temperature, and out-of-distribution predictions for (Ca1-xLax)FeAs2 based on varying lanthanum content. The outcome of our study underscores the significance of systematic feature analysis and selection in enhancing predictive model performance, offering various advantages over models that rely primarily on algorithmic complexity. This research not only advances the field of superconductivity but also sets a precedent for the application of machine learning in materials science.

Development of message passing-based graph convolutional networks for classifying cancer pathology reports

BackgroundApplying graph convolutional networks (GCN) to the classification of free-form natural language texts leveraged by graph-of-words features (TextGCN) was studied and confirmed to be an effective means of describing complex natural language texts. However, the text classification models based on the TextGCN possess weaknesses in terms of memory consumption and model dissemination and distribution. In this paper, we present a fast message passing network (FastMPN), implementing a GCN with message passing architecture that provides versatility and flexibility by allowing trainable node embedding and edge weights, helping the GCN model find the better solution. We applied the FastMPN model to the task of clinical information extraction from cancer pathology reports, extracting the following six properties: main site, subsite, laterality, histology, behavior, and grade.ResultsWe evaluated the clinical task performance of the FastMPN models in terms of micro- and macro-averaged F1 scores. A comparison was performed with the multi-task convolutional neural network (MT-CNN) model. Results show that the FastMPN model is equivalent to or better than the MT-CNN.ConclusionsOur implementation revealed that our FastMPN model, which is based on the PyTorch platform, can train a large corpus (667,290 training samples) with 202,373 unique words in less than 3 minutes per epoch using one NVIDIA V100 hardware accelerator. Our experiments demonstrated that using this implementation, the clinical task performance scores of information extraction related to tumors from cancer pathology reports were highly competitive.

Machine Learning for the Prediction of Surgical Morbidity in Placenta Accreta Spectrum.

We sought to create a machine learning (ML) model to identify variables that would aid in the prediction of surgical morbidity in cases of placenta accreta spectrum (PAS). A multicenter analysis including all cases of PAS identified by pathology specimen confirmation, across five tertiary care perinatal centers in New York City from 2013 to 2022. We developed models to predict operative morbidity using 213 variables including demographics, obstetrical information, and limited prenatal imaging findings detailing placental location. Our primary outcome was prediction of a surgical morbidity composite defined as including any of the following: blood loss (>1,500 mL), transfusion, intensive care unit admission, vasopressor use, mechanical ventilation/intubation, and organ injury. A nested, stratified, cross-validation approach was used to tune model hyperparameters and estimate generalizability. Gradient boosted tree classifier models incorporated preprocessing steps of standard scaling for numerical variables and one-hot encoding for categorical variables. Model performance was evaluated using area under the receiver operating characteristic curve (AUC), positive and negative predictive values (PPV, NPV), and F1 score. Variable importance ranking was also determined. Among 401 PAS cases, 326 (81%) underwent hysterectomy. Of the 401 cases of PAS, 309 (77%) had at least one event defined as surgical morbidity. Our predictive model had an AUC of 0.79 (95% confidence interval: 0.69, 0.89), PPV 0.79, NPV 0.76, and F1 score of 0.88. The variables most predictive of surgical morbidity were completion of a hysterectomy, prepregnancy body mass index (BMI), absence of a second trimester ultrasound, socioeconomic status zip code, BMI at delivery, number of prenatal visits, and delivery time of day. By identifying social and obstetrical characteristics that increase patients' risk, ML models are useful in predicting PAS-related surgical morbidity. Utilizing ML could serve as a foundation for risk and complexity stratification in cases of PAS to optimize surgical planning. · ML models are useful models are useful in predicting PAS-related surgical morbidity.. · Optimal management for PAS remains unclear.. · Utilizing ML can serve as a foundation for risk and complexity stratification in cases of PAS..

MobilenetV2-RC: A lightweight network model for retinopathy classification in retinal OCT images

Abstract Retinopathy is an important ophthalmic disease that causes blindness in the elderly population. With the global old peoples increasing, the ophthalmic healthcare system is more and significant to pre-diagnosis. Optical Coherence Tomography (OCT) is considered the gold standard for ophthalmic treatment and diagnosis. OCT technologies and equipment continue to develop towards the intelligence and convenience for requirements of rapid diagnosis in the remote and poverty-stricken areas. We proposed an improved MobilenetV2 lightweight model for retinopathy classification (MobilenetV2-RC), which incorporates Spatial and Channel Reconstruction Convolution (SCConv) and the improved Convolutional Block Attention Module (CBAM) attention mechanism into the framework. Not only can it effectively limit feature redundancy to reduce model parameters, but also enhance the ability of feature representation to improve classification accuracy. The parameters of the proposed model are only 1.96 M with an overall accuracy of 98.96%, which is higher 3.32% than the original MobilenetV2. Compared with ResNet18, InceptionV3, and VGG16_BN, the overall accuracy is increased by 4.6%, 6.3%, and 3.9%, respectively. The test results of UCSD and Duke open-source datasets are more remarkable. Experimental results show that our proposed algorithm has strong reliability and generalization for the accurate classification of retinopathy, and a greater application prospect in the intelligent diagnosis of ophthalmology and mobile detection terminals.

Medical image analysis using improved SAM-Med2D: segmentation and classification perspectives

Recently emerged SAM-Med2D represents a state-of-the-art advancement in medical image segmentation. Through fine-tuning the Large Visual Model, Segment Anything Model (SAM), on extensive medical datasets, it has achieved impressive results in cross-modal medical image segmentation. However, its reliance on interactive prompts may restrict its applicability under specific conditions. To address this limitation, we introduce SAM-AutoMed, which achieves automatic segmentation of medical images by replacing the original prompt encoder with an improved MobileNet v3 backbone. The performance on multiple datasets surpasses both SAM and SAM-Med2D. Current enhancements on the Large Visual Model SAM lack applications in the field of medical image classification. Therefore, we introduce SAM-MedCls, which combines the encoder of SAM-Med2D with our designed attention modules to construct an end-to-end medical image classification model. It performs well on datasets of various modalities, even achieving state-of-the-art results, indicating its potential to become a universal model for medical image classification.

Alzheimer’s disease (AD) classification using MRI: A deep ensemble model with modified local pattern feature set

The Alzheimer disease (AD) is a neurologic brain condition, which affects the cells in the brain and eventually renders a patient incapable of performing routine daily tasks. Due to the outstanding spatial clarity, high access, and strong contrast, MRI has been utilized in analyses pertaining to AD. This work develops an AD classification model using MRI images. Here, preprocessing is done by the Gabor filter. Subsequently, the Improved U-net segmentation model is employed for image segmentation. The features extracted comprises of modified LGXP features, LTP features, and LBP features as well. Finally, the Deep ensemble classifier (DEC) model is proposed for AD classification which combines classifiers such as RNN, DBN, and Deep Maxout Network (DMN). For enhancing the efficiency for classification of AD, the optimal weight of DMN is adjusted using the Self Customized BWO (SC-BWO) model. The outputs from DEC are averaged and the final result is obtained. Finally, the analysis of dice, Jaccard scores is performed to show the betterment of the SC-BWO scheme.

Cross-Domain Image Classification via Multimodal Augmentations and Bacterial Foraging Optimization

Identification of diseases from body scans requires design of pre-processing, filtering, segmentation, feature representation, classification & post-processing operations. Existing deep learning-based classification models use context-specific segmentation followed by convolutions or transformer-based classification techniques. However, the majority of these models do not perform correlative analysis between different disease types. This limitation restricts their scalability and applicability in clinical scenarios. To address these limitations, this research work proposed IMAC-MOC, a model that integrates multimodal augmentations to enhance performance in cross-domain and multi-organ image classification. The proposed model initially collects large-scale scans at organ-level like MRI, CT scan, Kidney Scans, etc. and represents them via multimodal feature sets. These feature maps are combined and processed via a Bacterial Foraging Optimizer (BFO) which assists in identification of high variance inter-class feature sets.

Automated Surgical Wound Classification for Intelligent Emergency Care Applications

Developing AI-powered solutions for emergency care is critical to improving patient outcomes. This study presents an advanced AI model designed to accurately classify surgical wounds to facilitate prompt and appropriate emergency response. We used two state-of-the-art image classification models, ResNeXt and Vision Transformer (ViT), to evaluate their effectiveness in classifying wound images. These models were selected based on preliminary evaluations of several popular models, and their superior performance metrics justified their final selection. The models were trained on a combined dataset of approximately 1,000 images from the Medetec and AZH(Advancing the Zenith of Healthcare) datasets. To improve classification accuracy, the dataset was preprocessed, including image resizing, flipping, normalization, and augmentation. The ViT model demonstrated superior performance, achieving an accuracy of 92.78%, precision of 94.89%, recall of 91.87%, and an F1 score of 92.44%. These results surpass those of existing studies such as Shenoy et al. (85.1% accuracy), Rostami et al. (68.7% accuracy), and Gao et al. (76.6% accuracy). Our proposed AI system not only accelerates life-saving first aid by providing timely and accurate wound classification but also enhances the skills of emergency responders through continuous learning and real-time feedback. By integrating AI into emergency service protocols, we aim to improve response times and collaboration among medical personnel, ultimately contributing to better patient survival rates.

Lung cancer classification model using convolutional neural network with feature ranking process

Lung cancer classification model using convolutional neural network with feature ranking process