Browse Items (16481 total)
Sort by:
-
Development vs. Rights A Case for Sustainable Development of Onge Tribes of Little Andaman
Human rights and environmental protections are often violated as a consequence of development activities. In addition to harming the environment, this increases the marginalisation of those who are already marginalised. The development paradigm that is based on the interests of the majority not only tends to retard the indigenous people but also renders them incapable of competing with the majority. For the indigenous people, development has always been a problem rather than a solution. Development initiatives under the umbrella of globalisation with a label of monotony, ignore the aspects of the diverse livelihoods of many indigenous peoples. The Niti Aayog proposed in its vision document, the Sustainable Development of Little Andaman, in 2021, that the island should be developed into a megacity by utilising its natural features and strategic location. The long-term objective is to develop the island into a major financial tourism hub that can rival Hong Kong and Singapore. This plan will, on the one hand, advance commerce, employment, and economic growth; on the other hand, environmental conservation issues will also arise. Concerns over this vision document have indeed been voiced by several academics, environmentalists, and conservationists due to issues with Onge indigenous rights, ecological fragility, and earthquake and tsunami susceptibility. In this context, the research article aims to study and analyse the proposed megacity project and its impact on the rights of Onge tribes and the environment. Sahana Florence and Achyutananda Mishra, 2024. -
She Shores A Study on the Lives, Challenges and Resilience of Women of the Koli Fishing Community in Mumbai
This study delves into the lives of women from the Koli fishing community in Mumbai, aiming to illuminate their unique life experiences and the daily struggles that often remain hidden beneath their prosperous facade. It endeavours to examine their agency and adaptive strategies employed to navigate these challenges. The research was conducted in Pachubandar, Vasai, located in the western suburbs of Mumbai, which stands as one of the prominent Koli settlements in the city. Employing a qualitative research approach coupled with an exploratory research design, the study engaged ten participants, comprising seven Koli women and three key informants from the community. Additionally, an observational analysis of four retail and wholesale fish markets in Mumbai was conducted to gain insight into the working conditions of Koli fisherwomen. This study adopts a gender-focused perspective to scrutinise the contextual vulnerabilities that shape the lives of Koli women. It underscores the paradox wherein, despite playing a pivotal role in sustaining both their families and the traditional fishing occupation, their contributions often go unnoticed. The Koli women face severe deprivation due to their limited access to property and decision-making authority. They find themselves entangled within traditional norms and patriarchal structures, which impede their access to essential assets and diverse livelihood resources. Although they significantly contribute to the fishery sector, their struggles, needs, and aspirations are frequently disregarded due to their lack of representation and involvement in decision-making bodies. The majority of these women work under precarious conditions, devoid of proper infrastructure, resources, and security. Furthermore, the evolving dynamics within the fishery sector, driven by rapid urbanisation and modernisation, have a profound impact on the lives and traditional livelihoods of Koli women. They now confront issues such as dwindling fish catches due to environmental degradation, heightened market competition, reduced livelihood spaces brought about by shifting urban and coastal landscapes, altered labour relations, and technological advancements. Consequently, they find themselves caught between the conflicting forces of tradition and modernity. The research also sheds light on the strategies devised by Koli women to resist and adapt to the uncertainties and challenges they encounter, ultimately safeguarding their livelihoods through self-organisation. The study emphasises the imperative to acknowledge their contributions as visible work and advocates for the incorporation of gender considerations when formulating policies and development strategies within the fisheries sector. MEGHNA ROY AND JYOTI SINGH, 2024. -
Perspectives on the Intersection of Gender, Customary Laws and Land Rights in India
For centuries, tribal communities in India have maintained distinct social and cultural identities, often with communal land ownership practices that were inclusive of women. The struggle of tribal women in India for land rights is a poignant manifestation of their fight against intersecting forms of oppression rooted in patriarchy, traditional power structures, and historical marginalisation. Given the existing background, this article discusses the intersection of property rights and gender relations in India, making a case for independent property rights for tribal women. It analyses the role of customary laws of inheritance in a legal pluralistic India and its conflict with positive law. The article also focuses on the role of the Indian judiciary in remedying the systemic discrimination against tribal women in India. It analyses the approach of the Indian courts in maintaining a balance between the autonomy granted to the tribes by the Indian Constitution and ensuring justice to women who are victims of such self-governance. Jyoti Singh and Kajori Bhatnagar, 2024. -
Large-Scale Proteomics Reveals New Candidate Biomarkers for Late-Onset Preeclampsia
BACKGROUND: Preeclampsia is classified as either a more severe early onset or a more prevalent late-onset form. Lower PlGF (placental growth factor) and increased sFlt-1 (fms-like tyrosine kinase-1) in maternal circulation are promising biomarkers, yet they lack specificity for preeclampsia. METHODS: We quantified ?7000 proteins in 673 samples collected from 89 patients with late-onset preeclampsia and 91 controls at T1 (1522), T2 (2230), and T3 (3042) weeks. Elastic net and random forest models were fitted and evaluated by cross-validation. Differential abundance analysis followed by functional profiling, was used to identify and interpret protein changes. RESULTS: An increase in protein differential abundance in late-onset preeclampsia was observed with advancing gestation, reaching 806 proteins at T3 related to angiogenesis, cell adhesion, and extracellular matrix remodeling. FAAH2 (fatty acid amide hydrolase 2), SIGLEC6 (sialic acid-binding Ig-like lectin-6), IL17RC (interleukin-17 receptor C), HTRA1 (serine protease), sFlt-1, and 47 other proteins dysregulated at T3 were validated in a reanalysis of a ?5000 protein Norwegian data set. Random forest models with 20 proteins showed high accuracy at T3 (area under the curve [AUC], 0.83 [0.770.89], sensitivity 59%) even in cases not yet diagnosed at sampling (n=31, AUC, 0.80 [0.710.90], sensitivity 58%), outperforming sFlt-1 and PlGF. Moderate accuracy was obtained at T1 (AUC, 0.63 [0.540.72], sensitivity 33%) and T2 (AUC, 0.59 [0.500.68], sensitivity 17%). Combining maternal characteristics and obstetric history with proteomics data increased accuracy at T1 (AUC, 0.68 [0.590.77], sensitivity 28%), T2 (AUC, 0.68 [0.600.77], sensitivity 31%), and T3 (AUC, 0.87 [0.810.92], sensitivity 69%). CONCLUSIONS: The findings confirm the involvement of abnormal trophoblast invasion, angiogenesis, and extracellular matrix remodeling in late-onset preeclampsia, while highlighting new protein alterations consistent across diverse cohorts. 2025 American Heart Association, Inc. -
In Silico Screening of Medicinal Plant-Derived Compounds Against Spodoptera litura
Spodoptera litura (Lepidoptera: Noctuidae) is a major agricultural pest in the Asia-Pacific region, causing significant crop damage. Current pest control strategies heav-ily rely on chemical pesticides, leading to environmental concerns and rapid resistance development. Molecular docking and molecular dynamics studies were used to investigate bio-compounds from three medicinal plants-Vitex negundo, Artemisia nilagirica, and Portulaca oleraceaas potential eco-friendly pest management al-ternatives. Gas chromatography-mass spectrometry (GC-MS) analysis identified 28 phytochemicals, of which 14 conformed to Lipinskis Rule of Five, which were selected as ligands. Molecular docking simulations were conducted to evaluate ligand interactions with four key target proteins in Spodoptera litura: acetylcholinesterase (AChE), carboxylesterase (CES), ecdysone receptor (EcR) and juvenile hormone (JH). Among the tested compounds, oxalic acid, 6-ethyloct-3-yl hexyl ester, and (11Z)-13-methyl-11-tetradecenyl acetate exhibited the highest binding affinities (-8.4 to-6.5 kcal/mol), suggesting their potential as inhibitory agents. Normal mode analysis (NMA) revealed low eigenvalues of the complexes, ranging from 9.6992260-5 to 3.0715890-4, indicating flexibility and requiring minimal energy for conformational changes. Deformability was highest in hinge regions, while var-iance analysis confirmed inverse proportionality across the complexes. The B-factor graph highlighted stable mobility and the root mean square (RMS) of the 3D con-former structures. Elastic network graphs displayed residue interactions as dots, with darker grey areas signifying greater stiffness. ADME/T analysis showed that these compounds possess favorable pharmacokinetic properties, including efficient ab-sorption and metabolism, while exhibiting no significant risks of mutagenicity or cardiotoxicity. These findings further support the suitability of Oxalic acid, 6-ethyloct-3-yl hexyl ester, and (11Z)-13-Methyl-11-tetradecenyl acetate as promising candidates for advancing sustainable and eco-friendly pest management approaches. Furthermore, the potential of identified plant-derived compounds as novel biopesti-cides contributes to sustainable and environmentally responsible pest management strategies. 2025, Brawijaya University. All rights reserved. -
Semantic segmentation for data validation in unmanned robotic vehicles
Semantic segmentation is a vital aspect of computer vision, widely used in fields such as autonomous driving, medical imaging, and industrial automation. Maintaining high-quality datasets is crucial for enhancing model accuracy and minimizing real-world errors. This paper focuses on developing a comprehensive data validation pipeline for semantic segmentation using OpenCV. The proposed framework integrates automated integrity checks, preprocessing techniques, and consistency verification to manage large-scale datasets effectively. Key validation processes include image quality assessment (detection of blurriness and noise), verification of annotation accuracy, class distribution analysis, and identification of anomalies. Additionally, OpenCV-powered preprocessing steps, such as image resizing, normalization, contrast optimization, and data augmentation, are applied to refine dataset quality for segmentation models. This paper also addresses scalability concerns associated with processing extensive datasets, introducing optimized batch handling and parallel validation techniques. By implementing a structured validation workflow, this research enhances the reliability, robustness, and overall effectiveness of semantic segmentation models, ensuring high-quality training data for deep learning applications. 2026, Intelektual Pustaka Media Utama. All rights reserved. -
Beliefs of secondary school teachers towards education for sustainable development: a statistical research
Educators are the architects of sustainable development (SD), transforming society and balancing development and sustainability. They enhance education for sustainable development (ESD) and societal transformation, driving innovative evolution and future-oriented development within the community. ESD, a millennium, and sustainable development goal (SDG), need to be implemented globally. Teachers are vital in transmitting knowledge, beliefs, and skills required for sustainability in the changing environment. This study examined secondary school teachers beliefs about ESD based on their professional qualifications, teaching experience, and position. The authors used a survey approach and collected the data using a belief assessment tool, i.e., the ESD beliefs scale. The respondents were 400 secondary school teachers in Kerala, India. The study used an item-based evaluation to achieve these objectives and calculated t-values, F-values, and percentages. The research findings indicated that teachers hold constructive opinions towards ESD. The positional status of teachers did not alter beliefs regarding ESD among secondary school teachers. In contrast, professional qualifications and years of teaching experience significantly influenced these ESD beliefs. The findings from this study enable education stakeholders to amend the current secondary education system for SD. 2026 Institute of Advanced Engineering and Science. All rights reserved. -
Development of interactive e-content to enhance listening skill and language comprehension among secondary school students
The present study aimed to develop interactive e-content, conduct expert validation, and examine the appropriate level. The researchers used a purposive sampling technique to select the sample of 100 secondary school students and 35 teachers from the Kerala state scheme. The researchers adopted the analysis, design, development, implementation, and evaluation (ADDIE) model to develop interactive e-content. The study employed two quantitative methods. Firstly, the study administered expert validation sheets to three content and two media experts to validate developed interactive e-content. The study utilized the percentage analysis to evaluate the results of the expert validation sheets. Secondly, the study administered a survey questionnaire to 100 secondary school students and 35 teachers to examine the appropriate level of interactive e-content. The study employed the correlation method to analyze the questionnaire results, examining the strength and direction of relationships between variables. The average score of content expert validation is 95.5% and media expert validation is 91.5% confirm that the developed interactive e-content is highly valid and appropriate. A major challenge for the researchers was the insufficient internet speed in rural areas of Kerala. The study recommends that teachers have to develop interactive multimedia teaching-learning aids to improve listening, speaking, reading, and writing (LSRW) among students. 2026 Institute of Advanced Engineering and Science. All rights reserved. -
The educational accomplishments scale: development and validation in the context of education institutions
Educational institutions play a significant role in fostering academic growth and personal development. However, there is a lack of standardized tools to assess the impact of educational accomplishments (EA), particularly integrating dimensions such as quality, value-based, integrated, and culture-enhanced education. This paper aims to create and validate a measurement tool that assesses how EA impacts students and institutions to foster academic growth, personal development, and institutional effectiveness, contributing to the overall quality of education. The data was collected from 120 participants, including religious heads, directors, principals, and coordinators of ten schools run by a specific religious congregation. The study implemented a three-stage systematic procedure in the development of the scale. Stage one consisted of item generation, literature review, and expert judgment. The second stage validated the scale and was followed by an item analysis, principal component with varimax rotation (exploratory factor analysis) using Kaiser normalization on IBM SPSS 26. The third step resulted in the final reliability and validity of the scale. A final 19-item educational accomplishments scale (EAS) is psychometrically reliable and of potential use to policymakers globally, comparing student and teacher perceptions, especially with religious congregational affiliations. This scale can particularly be used by each institution to evaluate the EA and can also be used by other researchers for further research. 2025 Institute of Advanced Engineering and Science. All rights reserved. -
Reflective writing skills among pre service teachers: a scoping review
Reflection is a soul-searching process. It is an innate ability to delve down the memory lane to judge a reaction to a particular situation as right or wrong as a response. The positive reactions are reinforced and the ineffective negative ones are relinquished. Developing reflective skills among preservice teachers include regular reflective practice sessions. They have to painstakingly record all their reflections after the delivery of each lesson as part of their curriculum along with other reflective practice opportunities. This effort should lead to evolution of professional practitioner in the long run. Although, there are factors affecting its development, preservice teachers seem to do it more monotonously without much reflective learning. Their reflective writing skills are way behind the expected level. This study adopts the research design outline advocated by Arksey and OMalley. The study appraised the research studies conducted from 2015 to 2024 as a part of scoping review. The study throws light on the various aspects related to the teacher-trainees reflective writing skills. Future studies may focus on empirical validation of the reflective writing skills among preservice teachers. 2025 Institute of Advanced Engineering and Science. All rights reserved. -
The impact of leader motives in students: a systematic review
Leader motives elucidate the driving forces behind leadership behavior and decision-making, which are pivotal for understanding effective leadership dynamics across diverse contexts. In this context, the systematic literature review (SLR) analyzed leader motives among students, providing insights into the underlying drivers shaping leadership behaviors within educational environments. This paper aims to understand how leader motives impact student behavior, academic performance, and social dynamics within educational environments. Based on McClellands needs theory as a conceptual framework, the review examines students prevalence and manifestations of achievement, power, and affiliation motives. This study systematically reviewed 16 papers, scholarly databases, and pertinent literature published between 2007 and 2024. A preferred reporting items for systematic reviews and meta-analysis (PRISMA) method was used to report the items. The findings underscore the importance of nurturing leader motives in educational settings, which contribute to positive student outcomes and foster leadership development through the lens of need theory. This study contributes to understanding how leader motives can elevate leadership behaviors and outcomes, offering valuable insights for policymakers and academic leaders aiming to enhance educational quality. 2025, Institute of Advanced Engineering and Science. All rights reserved. -
Opportunities and challenges while conducting field trips to the museum: a narrative review
The museum visit field trip engages and motivates the children in various activities. Field trips to the museum provide the students with a constructivist and experiential learning environment as they construct knowledge through observing the artifacts. The present study describes the possible opportunities and challenges for school children while conducting field trips to the museum. The study employed a narrative review technique to address the research question raised. The study selected the literature reviews from 2012-2023, including studies on field trips to the museum for the academic engagement of school children. The data includes 50 peer-reviewed journal articles categorized into five categories: students overall development, experiential learning opportunities, the museum as a resource, the role of teachers, the school, and museum authority. Results revealed that the museum is a resource for learning and is perfect for improving students cognitive and affective development towards the various school subjects and helping them enhance their participatory learning opportunities. However, teachers knowledge, infrastructure, parental consent, and legitimization with the school authority are some challenges in conducting museum visit field trips. Future research may focus on conducting empirical studies, which include school-museum collaboration, to enhance the horizon of school and community knowledge. 2025, Institute of Advanced Engineering and Science. All rights reserved. -
Lung cancer prediction with advanced graph neural networks
This research aims to enhance lung cancer prediction using advanced machine learning techniques. The major finding is that integrating graph convolutional networks (GCNs) with graph attention networks (GATs) significantly improves predictive accuracy. The problem addressed is the need for early and accurate detection of lung cancer, leveraging a dataset from Kaggle's "Lung Cancer Prediction Dataset," which includes 309 instances and 16 attributes. The proposed A-GCN with GAT model is meticulously engineered with multiple layers and hidden units, optimized through hyperparameter adjustments, early stopping mechanisms, and Adam optimization techniques. Experimental results demonstrate the model's superior performance, achieving an accuracy of 0.9454, precision of 0.9213, recall of 0.9743, and an F1 score of 0.9482. These findings highlight the model's efficacy in capturing intricate patterns within patient data, facilitating early interventions and personalized treatment plans. This research underscores the potential of graph-based methodologies in medical research, particularly for lung cancer prediction, ultimately aiming to improve patient outcomes and survival rates through proactive healthcare interventions. 2025 Institute of Advanced Engineering and Science. All rights reserved. -
A novel AI model for the extraction and prediction of Alzheimer disease from electronic health record
Dark data is an emerging concept, with its existence, identification, and utilization being key areas of research. This study examines various aspects and impacts of dark data in the healthcare domain and designs a model to extract essential clinical parameters for Alzheimer's from electronic health records (EHR). The novelty of dark data lies in its significant impact across sectors. In healthcare, even the smallest data points are crucial for diagnosis, prediction, and treatment. Thus, identifying and extracting dark data from medical data corpora enhances decision-making. In this research, a natural language processing (NLP) model is employed to extract clinical information related to Alzheimer's disease, and a machine learning algorithm is used for prediction. Named entity recognition (NER) with SpaCy is utilized to extract clinical departments from doctors' descriptions stored in EHRs. This NER model is trained on custom data containing processed EHR text and associated entity annotations. The extracted clinical departments can then be used for future Alzheimer's diagnosis via support vector machine (SVM) algorithms. Results show improved accuracy with the use of extracted dark data, highlighting its importance in predicting Alzheimer's disease. This research also explores the presence of dark data in various domains and proposes a dark data extraction model for the clinical domain using NLP. 2025 Institute of Advanced Engineering and Science. All rights reserved. -
An enhanced predictive modelling framework for highly accurate non-alcoholic fatty liver disease forecasting
Non-alcoholic fatty liver disease (NAFLD) is a chronic medical ailment characterized by accumulation of excessive fat in the liver of non-alcoholic patients. In absence of any early visible indications, application of machine learning based predictive techniques for early prediction of NAFLD are quite beneficial. The objective of this paper is to present a complete framework for guided development of varied predictive machine learning models and predict NAFLD disease with high accuracy. The framework employs stepby-step data quality enhancement to medical data such as cleaning, normalization, data upscaling using SMOTE (for handling class imbalances) and correlation analysis-based feature selection to predict NAFLD with high accuracy using only clinically recorded identifiers. Comprehensive comparative analysis of prediction results of seven machine learning predictive models is done using unprocessed as well as quality enhanced data. As per the observed results, XGBoost, random forest and neural network machine learning models reported significantly higher accuracies with improved AUC and ROC values using preprocessed data in contrast to unprocessed data. The prediction results are also assessed on various quality metrics such as accuracy, f1-score, precision, and recall significantly support the need for presented methodologies for qualitative NAFLD prediction modelling. 2025 Institute of Advanced Engineering and Science. All rights reserved. -
Deep learning ensembles for lung cancer detection in thoracic CT scans leveraging generative adversarial network technology
Effective treatment of lung cancer depends on early and accurate detection, which continues to be a major cause of cancer-related fatalities globally. Conventional diagnostic techniques are useful, but their efficacy in handling large amounts of thoracic computed tomography (CT) scan data is limited by their time-consuming nature and susceptibility to human error. The research here suggests a new deep learning model that integrates generative adversarial networks (GANs) for data improvement with a sophisticated ensemble approach to classification. GANs are employed to generate realistic synthetic CT images, addressing the challenges of limited datasets. The backbone of the proposed approach is a consensus-guided adaptive blending (CGAB) ensemble model that learns to dynamically combine the predictions of three top-performing convolutional neural networks (CNNs): ResNet-152, DenseNet-169, and EfficientNet-B7. The CGAB model improves prediction accuracy through model contribution weighting based on confidence scores and inter-model consensus, while a conflict-resolving auxiliary decision model is used. The approach was tested using the lung image database consortium and the image database resource initiative (LIDC-IDRI) dataset with a detection rate of 97.35, surpassing single-model and traditional ensemble methods. The current work provides a solid and scalable approach to lung cancer detection with better generalization, increased diagnostic consistency, and applicability for clinical use. This is an open access article under the CC BY-SA license. -
NN-SVM: a hybrid neural networksupport vector machine framework for accurate pneumonia detection from chest X-rays
We present neural network (NN)-support vector machine (SVM), hybrid NN-SVM framework for three-class pneumonia detection (normal, bacterial, and viral) from chest X-rays (CXRs). Pretrained NN backbone is fine-tuned for radiographic textures; global average pooling (GAP) yields embeddings that feed calibrated radial basis function (RBF)-SVM. Standardized preprocessing (resize, normalization) and class-aware augmentation are applied. We report accuracy, precision, recall, F1-score, area under the curve (AUC), confusion matrices, and per-class receiver operating characteristic (ROC). Statistical significance is assessed via DeLong (AUC), McNemar (accuracy), and paired bootstrap (F1-score). Gradient-weighted class activation mapping (grad-CAM) supports interpretability; external validation and domain adaptation (batch normalization re-estimation and temperature scaling) assess robustness. NN-SVM attains 97.46% accuracy with strong macro-F1 and AUC. Compared with SoftMax head, SVM improves margin separation and calibration. We present NN-SVM, hybrid deep learning approach that combines transfer-learned convolutional neural networks (CNNs) with SVM classifier to automatically diagnose pneumonia from CXRs into three clinically relevant categories: viral pneumonia, bacterial pneumonia, and normal. We use pre-trained CNN to extract robust image embeddings after standardized preprocessing (resizing and normalization) and train RBF-kernel SVM on resulting features. Performance is evaluated with accuracy, precision, recall, F1-score, and confusion matrices. On labeled CXR dataset, NN-SVM achieves 97.46% accuracy, demonstrating strong diagnostic capability that can reduce radiologist burden and support timely clinical decision-making. This is an open access article under the CC BY-SA license. https://creativecommons.org/licenses/by-sa/4.0/ -
Attribute optimization to improve breast cancer prediction using machine learning techniques
Breast cancer (BC) arises when cells grow out of control. It affects women more than men. Seeking cancer treatment can be both costly and time-consuming, with test results spanning from a few hours to several weeks. The duration of these tests depends on the number of attributes within the dataset. This research paper endeavors to optimize the dataset attributes and find the accuracy of the optimized dataset. The primary goal is to reduce features using recursive feature elimination to minimize the time taken for the test result. This work discusses the machine learning technique and the random forest (RF) algorithm, which helps determine the parameter accuracy on the Wisconsin BC diagnostic dataset. The method achieves an accuracy of 96.49% with only eighteen attributes. It has aided the healthcare industry in finding BC in less time and improving the treatment. Copyright (c) 2026 Peddireddy Venkateswara Reddy, Alaguchamy Parivazhagan. This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. -
Intelligent self-organizing microservice composition using hybrid learning for neonatal ward
This research presents an innovative self-organizing microservice composition model specifically tailored for dynamic and time-sensitive healthcare environments such as neonatal intensive care units (NICU). A hybrid machine learning classifier detects neonatal conditions and assigns treatment plans based on real-time vitals. The composition process is guided by a deep learning agent that combines unsupervised and reinforcement learning to develop intelligent bonding strategies. Microservices act as autonomous agents, supporting decentralized service choreography within the self-organizing framework. The bonding strategies of direct bonding and shared bonding are implemented for single conditions and coexisting conditions, respectively. The simulation results are based on actual NICU data, demonstrating the ability of the model to dynamically compose services while ensuring optimal resource utilization. The model demonstrates an adaptive and dynamic composition through emergence and continuous learning for changing clinical conditions, and demonstrates emergent behavior through reinforcement learning. The model's predictive capabilities enable anticipatory service loading, providing context-aware treatment in critical healthcare scenarios. This self-organizing architecture model offers a scalable and robust solution for autonomous, decentralized service choreography in critical healthcare environments. This is an open access article under the CC BY-SA license https://creativecommons.org/licenses/by-sa/4.0/. -
Pneumonia classification from chest X-rays using significant feature selection and machine learning
The chest X-ray images of normal lungs differ only subtly from those of lungs with pneumonia, making image-based diagnosis highly challenging. To address this issue, we developed a machine learning (ML)-based, lightweight, end-to-end Python package that processes chest X-ray images, implements robust feature selection methods, and classifies the images using various algorithms. While many studies have focused on improving classification accuracy using newer methods, few have addressed the interpretability of the extracted features or the growing computational demands of complex models. We used four publicly available datasets and extracted first-order, textural, and transform-based radiomic features to test our package. Features were selected using the Shapley additive explanations (SHAP) combined with recursive feature elimination (RFE) and stability selection algorithms. Our final solution contains a method that extracts a finite set of features identified by stability selection and feeds them as inputs into classical ML algorithms. Our model achieved 98% accuracy on the primary dataset, and 97%1, 96%2, and 94%2% accuracy on the other three datasets. Our approach is fast, self-contained, and requires only an ideal set of features, making it suitable for resource-constrained clinical environments. This is an open access article under the CC BY-SA license. https://creativecommons.org/licenses/by-sa/4.0/