Multi-class Classification Model Research Articles

A Hybrid Multi-class Classification Model for the Detection of Leaf Disease using XGBoost and SVM

A Hybrid Multi-class Classification Model for the Detection of Leaf Disease using XGBoost and SVM

Taste sensation evaluation for an electronic tongue based on an optimized computational model of taste pathways

Effective evaluation of taste sensation can be achieved by analyzing electronic tongue (e-tongue) data. Research on evaluation of the taste sensation of an e-tongue by nerve conduction mechanisms is limited, which obviously affects performance evaluation of e-tongues. Therefore, in this paper, a method for evaluating the taste sensation of an e-tongue based on human taste conduction mechanisms, the computational model of taste pathways (CMTP), is proposed. However, the limited physiological basis of the CMTP parameters influences the evaluation results. To achieve excellent evaluation performance, a parameter optimization algorithm with Hebbian and habituation learning rules is used to optimize the CMTP parameters. The effectiveness of the optimized results is demonstrated by improvement in the dynamic characteristics of the CMTP. Next, the optimized CMTP (OCMTP) is used for pattern recognition and sweetness evaluation of four taste substances. The results showed, first, that the dynamic characteristics (including 1/f characteristics and synchronization) of the OCMTP are improved, and the bionics of the OCMTP is enhanced. The optimized results are effective. Second, compared with the recognition results for the four taste substances by the unoptimized CMTP (UCMTP), signal preprocessing methods and multiclass classification models, the best classification accuracy (95.38%), the best kappa coefficient (93.83%) and the best F 1-score (96.10%) are achieved by the OCMTP. Finally, compared with the sweetness evaluation results of the UCMTP, signal preprocessing methods and multiple evaluation models, the best evaluation performance, including a root-mean-square error of 0.1643 and R 2 of 0.9785, is obtained using the OCMTP. In conclusion, effective evaluation of taste sensation can be achieved by the OCMTP and an e-tongue.

Predicting Chemical Carcinogens Using a Hybrid Neural Network Deep Learning Method.

Determining environmental chemical carcinogenicity is urgently needed as humans are increasingly exposed to these chemicals. In this study, we developed a hybrid neural network (HNN) method called HNN-Cancer to predict potential carcinogens of real-life chemicals. The HNN-Cancer included a new SMILES feature representation method by modifying our previous 3D array representation of 1D SMILES simulated by the convolutional neural network (CNN). We developed binary classification, multiclass classification, and regression models based on diverse non-congeneric chemicals. Along with the HNN-Cancer model, we developed models based on the random forest (RF), bootstrap aggregating (Bagging), and adaptive boosting (AdaBoost) methods for binary and multiclass classification. We developed regression models using HNN-Cancer, RF, support vector regressor (SVR), gradient boosting (GB), kernel ridge (KR), decision tree with AdaBoost (DT), KNeighbors (KN), and a consensus method. The performance of the models for all classifications was assessed using various statistical metrics. The accuracy of the HNN-Cancer, RF, and Bagging models were 74%, and their AUC was ~0.81 for binary classification models developed with 7994 chemicals. The sensitivity was 79.5% and the specificity was 67.3% for the HNN-Cancer, which outperforms the other methods. In the case of multiclass classification models with 1618 chemicals, we obtained the optimal accuracy of 70% with an AUC 0.7 for HNN-Cancer, RF, Bagging, and AdaBoost, respectively. In the case of regression models, the correlation coefficient (R) was around 0.62 for HNN-Cancer and RF higher than the SVM, GB, KR, DTBoost, and NN machine learning methods. Overall, the HNN-Cancer performed better for the majority of the known carcinogen experimental datasets. Further, the predictive performance of HNN-Cancer on diverse chemicals is comparable to the literature-reported models that included similar and less diverse molecules. Our HNN-Cancer could be used in identifying potentially carcinogenic chemicals for a wide variety of chemical classes.

CARE: coherent actionable recourse based on sound counterfactual explanations

Counterfactual explanation (CE) is a popular post hoc interpretability approach that explains how to obtain an alternative outcome from a machine learning model by specifying minimum changes in the input. In line with this context, when the model’s inputs represent actual individuals, actionable recourse (AR) refers to a personalized CE that prescribes feasible changes according to an individual’s preferences. Hence, the quality of ARs highly depends on the soundness of underlying CEs and the proper incorporation of user preferences. To generate sound CEs, several data-level properties, such as proximity and connectedness, should be taken into account. Meanwhile, personalizing explanations demands fulfilling important user-level requirements, like coherency and actionability. The main obstacles to inclusive consideration of the stated properties are their associated modeling and computational complexity as well as the lack of a systematic approach for making a rigorous trade-off between them based on their importance. This paper introduces CARE, an explanation framework that addresses these challenges by formulating the properties as intuitive and computationally efficient objective functions, organized in a modular hierarchy and optimized using a multi-objective optimization algorithm. The devised modular hierarchy enables the arbitration and aggregation of various properties as well as the generation of CEs and AR by choice. CARE involves individuals through a flexible language for defining preferences, facilitates their choice by recommending multiple ARs, and guides their action steps toward their desired outcome. CARE is a model-agnostic approach for explaining any multi-class classification and regression model in mixed-feature tabular settings. We demonstrate the efficacy of our framework through several validation and benchmark experiments on standard data sets and black box models.

A Multiclass Classification Model for Tooth Removal Procedures

Surprisingly little is known about tooth removal procedures. This might be due to the difficulty of gaining reliable data on these procedures. To improve our understanding of these procedures, machine learning techniques were used to design a multiclass classification model of tooth removal based on force, torque, and movement data recorded during tooth removal. A measurement setup consisting of, among others, robot technology was used to gather high-quality data on forces, torques, and movement in clinically relevant dimensions. Fresh-frozen cadavers were used to match the clinical situation as closely as possible. Clinically interpretable variables or “features” were engineered and feature selection took place to process the data. A Gaussian naive Bayes model was trained to classify tooth removal procedures. Data of 110 successful tooth removal experiments were available to train the model. Out of 75 clinically designed features, 33 were selected for the classification model. The overall accuracy of the classification model in 4 random subsamples of data was 86% in the training set and 54% in the test set. In 95% and 88%, respectively, the model correctly classified the (upper or lower) jaw and either the right class or a class of neighboring teeth. This article discusses the design and performance of a multiclass classification model for tooth removal. Despite the relatively small data set, the quality of the data was sufficient to develop a first model with reasonable performance. The results of the feature engineering, selection process, and the classification model itself can be considered a strong first step toward a better understanding of these complex procedures. It has the potential to aid in the development of evidence-based educational material and clinical guidelines in the near future.

High-dimensional multinomial multiclass severity scoring of COVID-19 pneumonia using CT radiomics features and machine learning algorithms

We aimed to construct a prediction model based on computed tomography (CT) radiomics features to classify COVID-19 patients into severe-, moderate-, mild-, and non-pneumonic. A total of 1110 patients were studied from a publicly available dataset with 4-class severity scoring performed by a radiologist (based on CT images and clinical features). The entire lungs were segmented and followed by resizing, bin discretization and radiomic features extraction. We utilized two feature selection algorithms, namely bagging random forest (BRF) and multivariate adaptive regression splines (MARS), each coupled to a classifier, namely multinomial logistic regression (MLR), to construct multiclass classification models. The dataset was divided into 50% (555 samples), 20% (223 samples), and 30% (332 samples) for training, validation, and untouched test datasets, respectively. Subsequently, nested cross-validation was performed on train/validation to select the features and tune the models. All predictive power indices were reported based on the testing set. The performance of multi-class models was assessed using precision, recall, F1-score, and accuracy based on the 4 × 4 confusion matrices. In addition, the areas under the receiver operating characteristic curves (AUCs) for multi-class classifications were calculated and compared for both models. Using BRF, 23 radiomic features were selected, 11 from first-order, 9 from GLCM, 1 GLRLM, 1 from GLDM, and 1 from shape. Ten features were selected using the MARS algorithm, namely 3 from first-order, 1 from GLDM, 1 from GLRLM, 1 from GLSZM, 1 from shape, and 3 from GLCM features. The mean absolute deviation, skewness, and variance from first-order and flatness from shape, and cluster prominence from GLCM features and Gray Level Non Uniformity Normalize from GLRLM were selected by both BRF and MARS algorithms. All selected features by BRF or MARS were significantly associated with four-class outcomes as assessed within MLR (All p values < 0.05). BRF + MLR and MARS + MLR resulted in pseudo-R2 prediction performances of 0.305 and 0.253, respectively. Meanwhile, there was a significant difference between the feature selection models when using a likelihood ratio test (p value = 0.046). Based on confusion matrices for BRF + MLR and MARS + MLR algorithms, the precision was 0.856 and 0.728, the recall was 0.852 and 0.722, whereas the accuracy was 0.921 and 0.861, respectively. AUCs (95% CI) for multi-class classification were 0.846 (0.805–0.887) and 0.807 (0.752–0.861) for BRF + MLR and MARS + MLR algorithms, respectively. Our models based on the utilization of radiomic features, coupled with machine learning were able to accurately classify patients according to the severity of pneumonia, thus highlighting the potential of this emerging paradigm in the prognostication and management of COVID-19 patients.

The utility of texture analysis of kidney MRI for evaluating renal dysfunction with multiclass classification model

We evaluated a multiclass classification model to predict estimated glomerular filtration rate (eGFR) groups in chronic kidney disease (CKD) patients using magnetic resonance imaging (MRI) texture analysis (TA). We identified 166 CKD patients who underwent MRI comprising Dixon-based T1-weighted in-phase (IP)/opposed-phase (OP)/water-only (WO) images, apparent diffusion coefficient (ADC) maps, and T2* maps. The patients were divided into severe, moderate, and control groups based on eGFR borderlines of 30 and 60 mL/min/1.73 m2. After extracting 93 texture features (TFs), dimension reduction was performed using inter-observer reproducibility analysis and sequential feature selection (SFS) algorithm. Models were created using linear discriminant analysis (LDA); support vector machine (SVM) with linear, rbf, and sigmoid kernels; decision tree (DT); and random forest (RF) classifiers, with synthetic minority oversampling technique (SMOTE). Models underwent 100-time repeat nested cross-validation. Overall performances of our classification models were modest, and TA based on T1-weighted IP/OP/WO images provided better performance than those based on ADC and T2* maps. The most favorable result was observed in the T1-weighted WO image using RF classifier and the combination model was derived from all T1-weighted images using SVM classifier with rbf kernel. Among the selected TFs, total energy and energy had weak correlations with eGFR.

Adaptive Thresholding of CNN Features for Maize Leaf Disease Classification and Severity Estimation

Convolutional neural networks (CNNs) are the gold standard in the machine learning (ML) community. As a result, most of the recent studies have relied on CNNs, which have achieved higher accuracies compared with traditional machine learning approaches. From prior research, we learned that multi-class image classification models can solve leaf disease identification problems, and multi-label image classification models can solve leaf disease quantification problems (severity analysis). Historically, maize leaf disease severity analysis or quantification has always relied on domain knowledge—that is, experts evaluate the images and train the CNN models based on their knowledge. Here, we propose a unique system that achieves the same objective while excluding input from specialists. This avoids bias and does not rely on a “human in the loop model” for disease quantification. The advantages of the proposed system are many. Notably, the conventional system of maize leaf disease quantification is labor intensive, time-consuming and prone to errors since it lacks standardized diagnosis guidelines. In this work, we present an approach to quantify maize leaf disease based on adaptive thresholding. The experimental work of our study is in three parts. First, we train a wide variety of well-known deep learning models for maize leaf disease classification, then we compare the performance of the deep learning models and finally extract the class activation heatmaps from the prediction layers of the CNN models. Second, we develop an adaptive thresholding technique that automatically extracts the regions of interest from the class activation maps without any prior knowledge. Lastly, we use these regions of interest to estimate image leaf disease severity. Experimental results show that transfer learning approaches can classify maize leaf diseases with up to 99% accuracy. With a high quantification accuracy, our proposed adaptive thresholding method for CNN class activation maps can be a valuable contribution to quantifying maize leaf diseases without relying on domain knowledge.

A novel multi-class classification model for schizophrenia, bipolar disorder and healthy controls using comprehensive transcriptomic data

Two common psychiatric disorders, schizophrenia (SCZ) and bipolar disorder (BP), confer lifelong disability and collectively affect 2% of the world population. Because the diagnosis of psychiatry is based only on symptoms, developing more effective methods for the diagnosis of psychiatric disorders is a major international public health priority. Furthermore, SCZ and BP overlap considerably in terms of symptoms and risk genes. Therefore, the clarity of the underlying etiology and pathology remains lacking for these two disorders. Although many studies have been conducted, a classification model with higher accuracy and consistency was found to still be necessary for accurate diagnoses of SCZ and BP. In this study, a comprehensive dataset was combined from five independent transcriptomic studies. This dataset comprised 120 patients with SCZ, 101 patients with BP, and 149 healthy subjects. The partial least squares discriminant analysis (PLS-DA) method was applied to identify the gene signature among multiple groups, and 341 differentially expressed genes (DEGs) were identified. Then, the disease relevance of these DEGs was systematically performed, including (α) the great disease relevance of the identified signature, (β) the hub genes of the protein-protein interaction network playing a key role in psychiatric disorders, and (γ) gene ontology terms and enriched pathways playing a key role in psychiatric disorders. Finally, a popular multi-class classifier, support vector machine (SVM), was applied to construct a novel multi-class classification model using the identified signature for SCZ and BP. Using the independent test sets, the classification capacity of this multi-class model was assessed, which showed this model had a strong classification ability.

Multi-Class Text Classification for Automated Team Communication Encoding

Team communication can be a window into team processes and the generation of team performance. Current methods of team communication encoding typically call for the manual application of a researcher derived coding scheme. Yet, such a coding approach can be resource and time intensive given the need for multiple coders and high reliability in the interpretation and assignment of codes. Such approaches on top of the time for audio data collection and processing may create difficulties in reaching the critical mass needed for methodological analyses in a useful time-frame. To enable faster, consistent team communication encoding, we evaluated different natural language processing techniques on team communication transcripts in assigning a coding scheme of process and purpose types to messages. We conducted a classifier selection of known multi-class text classification models, and a feature selection comparing word segmentation and parts-of-speech tagging. We were able to identify that the ensemble method using n-grams of parts-of-speech tags vectorized using TF-IDF Vectorizer and a composite classifier created using a pipeline of Standard Scalar Transformer with Decision Tree classifier was able to achieve an average accuracy of 91.6% for automatically assigning process codes and an average accuracy of 94.9% for message purpose codes. We discuss how researchers can use machine learning methods to move beyond the manual encoding of coding schemes to a more efficient and consistent approach.

Machine Learning Prediction of Hypoglycemia and Hyperglycemia From Electronic Health Records: Algorithm Development and Validation.

BackgroundAcute blood glucose (BG) decompensations (hypoglycemia and hyperglycemia) represent a frequent and significant risk for inpatients and adversely affect patient outcomes and safety. The increasing need for BG management in inpatients poses a high demand on clinical staff and health care systems in addition.ObjectiveThis study aimed to generate a broadly applicable multiclass classification model for predicting BG decompensation events from patients’ electronic health records to indicate where adjustments in patient monitoring and therapeutic interventions are required. This should allow for taking proactive measures before BG levels are derailed.MethodsA retrospective cohort study was conducted on patients who were hospitalized at a tertiary hospital in Bern, Switzerland. Using patient details and routine data from electronic health records, a multiclass prediction model for BG decompensation events (<3.9 mmol/L [hypoglycemia] or >10, >13.9, or >16.7 mmol/L [representing different degrees of hyperglycemia]) was generated based on a second-level ensemble of gradient-boosted binary trees.ResultsA total of 63,579 hospital admissions of 38,250 patients were included in this study. The multiclass prediction model reached specificities of 93.7%, 98.9%, and 93.9% and sensitivities of 67.1%, 59%, and 63.6% for the main categories of interest, which were nondecompensated cases, hypoglycemia, or hyperglycemia, respectively. The median prediction horizon was 7 hours and 4 hours for hypoglycemia and hyperglycemia, respectively.ConclusionsElectronic health records have the potential to reliably predict all types of BG decompensation. Readily available patient details and routine laboratory data can support the decisions for proactive interventions and thus help to reduce the detrimental health effects of hypoglycemia and hyperglycemia.

A Deep Learning Approach to Detect the Spoiled Fruits

Fruits are one of the vital sources of nutrients for the mankind and their life span is very less. The fruit spoilage may occur at various stages such as, at the harvest time, during transportation, during storage etc. Freshness is a parameter used for accessing the quality of the fruit. About 20% of the harvested fruits are spoiled due to many factors, before consumption by humans. The spoilage of one fruit has a direct impact on the neighboring fruits. It is also a one of the indicators that gives an estimation of number of days that a fruit can be preserved. Early identification of the spoilage helps in taking the appropriate measures for the removal of spoiled fruits from the whole lot. So that it helps in preventing the spread of spoilage to its adjacent fruits. Deep learning based technological advancements helps in automatically identifying the spoiled fruits. In this work, internal quality attributes of the fruit are not taken into consideration for spoilage detection, only the external attributes are considered. The supervised learning technique is employed for the freshness analysis of two different types of fruits, Apple and Banana. As the 2 varieties are involved, it is a multiclass classification model with 4 classes. One shot detection technique is employed to accurately classify among the good fruit and spoiled fruit. Few images in the dataset are obtained from the kaggle.com and the rest are self - captured images. The dataset is balanced to avoid the biasness in the model. The model is implemented using Yolov4 and tiny Yolov4 frame works. These are one shot detection techniques, can be used for real time deployment. The inferences were obtained on the real time images and video. Confusion matrix is tabulated the performance metrics such as accuracy, F1 Score and recall are discussed with respect to these two techniques.

Development of predictive modeling and deep learning classification of taxi trip tolls

Several studies discussed the predictive modeling of deep learning in different applications such as classifying tissue features from microstructural data, Crude Oil Prices, mechanical constitutive behavior of materials, microbiome data, and mineral prospectively. Commercial navigation includes a wealth of trip-related data, including distance, expected journey time, and tolls that may be encountered along the way. Using a classification algorithm, it is possible to extract drop-off and pickup locations from taxi trip data and estimate if the tour would incur tolls. In this work, let’s use the classification learner to create classification models, compare their performance, and export the findings for additional study. The workflow for the classification learner is the same as for the regression learner. The purpose is to make predictions based on fresh data in order to see how well the model performs with new data. To train the model, it’s critical to separate the data set. The combined training and validation data is next pre-processed, which involves tasks such as cleaning and developing new features skills. Once the data has been prepared, it’s time to begin the supervised machine learning process and test a number of ways to identify the best model, such as the type of model that should be used, the important features, and the best parameters of the model to find the best fit for the considered data. The results of analyzing different predictive multiclass classification models with taxi trip tolls show that it is possible to use a machine learning-based model when we like to avoid road tolls depending on historical data on taxi trip tolls. The outcome of this study can help to expect road tolls from the drop-off and pickup locations of a taxi data

Machine-learning in motor neuron diseases: Prospects and pitfalls.

Machine-learning in motor neuron diseases: Prospects and pitfalls.

Prediction Breast Molecular Typing of Invasive Ductal Carcinoma Based on Dynamic Contrast Enhancement Magnetic Resonance Imaging Radiomics Characteristics: A Feasibility Study.

ObjectiveTo investigate the feasibility of radiomics in predicting molecular subtype of breast invasive ductal carcinoma (IDC) based on dynamic contrast enhancement magnetic resonance imaging (DCE-MRI).MethodsA total of 303 cases with pathologically confirmed IDC from January 2018 to March 2021 were enrolled in this study, including 223 cases from Fudan University Shanghai Cancer Center (training/test set) and 80 cases from Shaoxing Central Hospital (validation set). All the cases were classified as HR+/Luminal, HER2-enriched, and TNBC according to immunohistochemistry. DCE-MRI original images were treated by semi-automated segmentation to initially extract original and wavelet-transformed radiomic features. The extended logistic regression with least absolute shrinkage and selection operator (LASSO) penalty was applied to identify the optimal radiomic features, which were then used to establish predictive models combined with significant clinical risk factors. Receiver operating characteristic curve (ROC), calibration curve, and decision curve analysis were adopted to evaluate the effectiveness and clinical benefit of the models established.ResultsOf the 223 cases from Fudan University Shanghai Cancer Center, HR+/Luminal cancers were diagnosed in 116 cases (52.02%), HER2-enriched in 71 cases (31.84%), and TNBC in 36 cases (16.14%). Based on the training set, 788 radiomic features were extracted in total and 8 optimal features were further identified, including 2 first-order features, 1 gray-level run length matrix (GLRLM), 4 gray-level co-occurrence matrices (GLCM), and 1 3D shape feature. Three multi-class classification models were constructed by extended logistic regression: clinical model (age, menopause, tumor location, Ki-67, histological grade, and lymph node metastasis), radiomic model, and combined model. The macro-average areas under the ROC curve (macro-AUC) for the three models were 0.71, 0.81, and 0.84 in the training set, 0.73, 0.81, and 0.84 in the test set, and 0.76, 0.82, and 0.83 in the validation set, respectively.ConclusionThe DCE-MRI-based radiomic features are significant biomarkers for distinguishing molecular subtypes of breast cancer noninvasively. Notably, the classification performance could be improved with the fusion analysis of multi-modal features.

Supervised Machine Learning-Based Decision Support for Signal Validation Classification.

IntroductionSignal validation in pharmacovigilance is the process of evaluating data to decide whether evidence is sufficient to justify further assessment of a detected signal. During the signal validation process, safety experts in our organization are required to review signals of disproportionate reporting (SDRs) and classify them into one of six predefined categories.ObjectiveThis experiment explored the extent to which predictive machine learning (ML) models can support the decision making of safety experts by accurately identifying the most appropriate predefined signal validation category.MethodsWe extracted cumulative data for six medicinal products, consisting of historic SDR validations and Individual Case Safety Reports, from the company’s safety database for training and testing of the ML model. We implemented a decision tree-based supervised multiclass classifier model termed Gradient Boosted Trees followed by a SHapley Additive exPlanations (SHAP) analysis to mitigate the “black box” effect of the ensemble model by identifying the key predicting features in the model. Following a retrospective analysis, a prospective experiment was conducted to test the model accuracy and user acceptance in a real-life setting.ResultsThe prediction accuracy of our ML model ranged from 83 to 86% over 3 months for the six medicinal products. The applicability of the model was confirmed by the company’s safety experts. Additionally, the systematic predictions provided valuable information to the safety experts and assisted them in reviewing the SDRs efficiently and consistently.ConclusionsThis experiment demonstrated that it is possible to train a multiclass classification model to accurately predict signal validation categories for SDRs. More importantly, the transparency of the predictions provided by the SHAP analysis led to high acceptance by the safety experts.

Effects of Oversampling Smote and Spectral Transformations in the Classification of Mango Cultivars Using Near-Infrared Spectroscopy

Near-Infrared spectroscopy (NIR) is a non-destructive analytical technique that can provide chemical and structural information on samples in a speedy and accurate time. NIR has a wavelength of 750-2500 nm. However, the absorbance bands of the NIR spectrum are often broad, non-specific, and overlapping. NIR spectrum analysis requires a multivariate method which is very subjective to noise arising from instrumentation. There is no standard protocol in modeling for classification and prediction using NIR spectra. Several models have been developed with and without pre-processing techniques. The SMOTE technique can improve the model to predict all class responses accurately. This research contributes to creating a multiclass classification model for grouping mango cultivars by finding the best pre-processing technique and using SMOTE oversampling. The results of the four test scenarios on the model's performance built using the Support Vector Machine (SVM) that the best model is obtained using spectral transformations with LSNV and CLIP operations with 100% accuracy, precision, and recall values. The Decision Tree (DT) has the performance results in 100% model was obtained by using spectral transformation with LSNV, CLIP and SAVGOL operations with parameters {'deriv_order': 0,1, 2, 'filter_win': 11, 13, 'poly_order': 3}. Using SMOTE has better accuracy than without pre-processing, with an accuracy of 92% on SVM and 94% on DT. In comparison, the combination of SMOTE and Spectral Transformation gives classification results for SVM and DT with the same accuracy of 96%, better than using SMOTE only.

An Intelligent Approach for Preserving the Privacy and Security of a Smart Home Based on IoT Using LogitBoost Techniques

Development and use of IoT devices have grown significantly in recent years. Many departments such as smart homes, smart healthcare, smart sports analysis, and different smart industries use IoT-based devices. In IoT devices, traffic is a very important part. IoT device traffic is distinct from traditional device traffic in various respects. In this study, 41 Internet-of-Things (IoT) devices were used. IoT devices provided 13 network traffic attributes to construct a multiclass classification model. Pre-processing techniques such as Normalization and Scaling of Dataset were used to pre-process the raw data acquired. Features can be extracted from text data using feature engineering algorithms. After stratification, the dataset contains 117,423 feature vectors utilized to develop the classification model further. Multiple performance metrics were used to demonstrate how well LogitBoost algorithms perform in this research. Using ensemble-based hybrid machine learning models to detect network anomalies in this research is an early step in developing an intrusion detection system (IDS). The main objective of this study is to detect attacks and anomalies in an IoT environment in a smart home. We have proposed a novel approach to developing LogitBoost algorithms, i.e., Logi-XGB, Logi-GBC, Logi-ABC, Logi-CBC, Logi-LGBM, and Logi-HGBC. After applying LogitBoost algorithms to the dataset for the classification, Logi-XGB scored 80.20% accuracy, and Logi-GBC scored 77.80% accuracy. Logi-ABC scored 80.33% accuracy. Logi-CBC scored the highest accuracy of 85.66%. Logi-LGBM and Logi-HGBC scored the same accuracy of 81.37%. Compared with previous LogitBoost algorithms implemented in previous studies, our proposed Logi-CBC has scored the highest accuracy on the given dataset.

Building Multiclass Classification Model of Logistic Regression and Decision Tree Using the Chi-Square Test for Variable Selection Method

The growth and development of children under five (toddlers) affect their health conditions. Each region uniquely identifies the main factors influencing the toddler's health condition. The status of toddlers is generally categorized into two classes, namely normal and abnormal, so it is often found that the condition of toddler status is in the form of multi-response variables. Combining the two binary classes' response variables will form a multiclass response variable requiring different model development techniques and performance measurements. This study aims to determine the main factors that affect toddlers' health conditions in Malang, Indonesia, build multiclass logistic regression and decision tree classification models, and measure the model's performance. The Chi-square test selected predictor features as the input of multiclass logistic regression and decision tree models. From the feature selection, four main factors influence the status of toddlers' health conditions in Malang: the mother's history of diabetes before pregnancy, the father's blood pressure, psychological condition, and drinking water quality. The decision tree model performs better than the logistic regression model on the various performance measures used.

LungNet22: A Fine-Tuned Model for Multiclass Classification and Prediction of Lung Disease Using X-ray Images.

In recent years, lung disease has increased manyfold, causing millions of casualties annually. To combat the crisis, an efficient, reliable, and affordable lung disease diagnosis technique has become indispensable. In this study, a multiclass classification of lung disease from frontal chest X-ray imaging using a fine-tuned CNN model is proposed. The classification is conducted on 10 disease classes of the lungs, namely COVID-19, Effusion, Tuberculosis, Pneumonia, Lung Opacity, Mass, Nodule, Pneumothorax, and Pulmonary Fibrosis, along with the Normal class. The dataset is a collective dataset gathered from multiple sources. After pre-processing and balancing the dataset with eight augmentation techniques, a total of 80,000 X-ray images were fed to the model for classification purposes. Initially, eight pre-trained CNN models, AlexNet, GoogLeNet, InceptionV3, MobileNetV2, VGG16, ResNet 50, DenseNet121, and EfficientNetB7, were employed on the dataset. Among these, the VGG16 achieved the highest accuracy at 92.95%. To further improve the classification accuracy, LungNet22 was constructed upon the primary structure of the VGG16 model. An ablation study was used in the work to determine the different hyper-parameters. Using the Adam Optimizer, the proposed model achieved a commendable accuracy of 98.89%. To verify the performance of the model, several performance matrices, including the ROC curve and the AUC values, were computed as well.