Learning Solutions Research Articles

Deep Learning-Based Detection of Android Malware using Graph Convolutional Networks (GCN)

The study is centered around identifying Android malware using deep learning methods through Graph Neural Networks (GNNs) and Graph Convolutional Networks (GCNs). With Android being widely used worldwide ensuring the security of released applications poses a challenge. Conventional malware detection techniques, like dynamic analysis have limitations in recognizing new malware types leading to a shift towards machine learning and deep learning solutions. The research introduces a malware detection system that employs GNNs particularly focusing on GCNs to analyze the relationships within an applications code by transforming APK files into graph formats. The system follows stages including data gathering, feature extraction, graph construction, model training and implementation. By concentrating on function call graphs the system proves effective in identifying software surpassing traditional machine learning methods in terms of accuracy, precision, recall and F1 score. The GCN based model shows enhancements over approaches with an accuracy rate of 95% compared to 89%, for traditional machine learning models. This progress highlights the potential of learning techniques in bolstering Android security. The system excels not in identifying software but also proves versatile, for different uses like screening apps, in stores and functioning as a standalone antivirus program.

Https://gurukuljournal.com/virtual-classroom/

Abstract— This research paper presents the design, implementation, and evaluation of a virtual classroom system developed using Java. With the increasing demand for remote learning solutions, the project aims to provide an interactive and collaborative platform for students and instructors to engage in online education. The paper discusses the system architecture, features, implementation details, and user experience evaluation, highlighting the potential of Java-based virtual classrooms in modern education. IndexTerms – Web-Based Application, Event Creation, Event tracking, and School Academic Calendar, MERN Stack.

Data-distribution-informed Nyström approximation for structured data using vector quantization-based landmark determination

We present an effective method for supervised landmark selection in sparse Nyström approximations of kernel matrices for structured data. Our approach transforms structured non-vectorial input data, like graphs or text, into a dissimilarity representation, facilitating the identification of data-distribution-informed landmarks through prototype-based learning. Experimental results indicate competitive approximation quality when compared to existing strategies, showcasing the advantageous impact of incorporating more information into the Nyström landmark selection process. This positions our method as an efficient and versatile solution for large-scale kernel learning.

PackMASNet: An information integration approach for quality inspection in industry 5.0

Industry 5.0 (I5.0) is increasingly governed by Operation Technology and Information Technology. It hosts Deep Learning (DL) solutions as a premier tool for the informatization of Quality Inspection (QI) of Mass Produced Products (MPP) and Mass Customized Products (MCP) at the on-premise edge devices. Besides, these ever-evolving MCPs in I5.0 require newer DL models to informatize customized quality parameters, mandating significant storage and computing resources, that is challenging to achieve at resource-constraint edge platform. This work proposes PackMASNet, a continual learning-based information integration scheme that sustainably incorporates the QI information of multiple MCPs in a single Deep Neural Network (DNN) under the following constraints: (i) resource-limited on-premise edge devices, (ii) no access to prior product data, and (iii) negligible deterioration of quality information of frequently manufactured MPPs. A lightweight Cosine-similarity threshold mechanism is incorporated to leverage the benefits of knowledge transfer (cloning) while avoiding its adverse impact in multi-varied MCPs. The scheme proposes DNN performance, storage, computation, data, and time efficient solutions for industries implementing the edge computing paradigm on I5.0. The efficacy of the proposed scheme is validated for the predictive QI of real-world use case in injection molding. The scheme preserves the frequently used information of MPP by 99.07%, saves computation cost by 132 times, reduces the DNN training time by 20.45%, and improves overall performance by 43.60% over the state-of-the-art schemes.

Inferring effective electrostatic interaction of charge-stabilized colloids from scattering using deep learning

An innovative strategy is presented that incorporates deep auto-encoder networks into a least-squares fitting framework to address the potential inversion problem in small-angle scattering. To evaluate the performance of the proposed approach, a detailed case study focusing on charged colloidal suspensions was carried out. The results clearly indicate that a deep learning solution offers a reliable and quantitative method for studying molecular interactions. The approach surpasses existing deterministic approaches with respect to both numerical accuracy and computational efficiency. Overall, this work demonstrates the potential of deep learning techniques in tackling complex problems in soft-matter structures and beyond.

TransKinect: a computer vision and machine learning clinical decision support system for automatic independent wheelchair transfer technique assessment

Background Physical and occupational therapists provide routine care for manual wheelchair users and are responsible for training and assessing the quality of transfers. These transfers can produce large loads on the upper extremity joints if improper sitting-pivot-technique is used. Methods to assess quality of transfers include the Transfer Assessment Instrument, a clinically validated tool derived from quantitative biomechanical features; however, adoption of this tool is low due to the complex usage requirements and speed of typical transfers. Objective The objective of this study is to develop and validate a computer vison and machine learning solution to better implement the Transfer Assessment Instrument in clinical settings. Methods The prototype system, TransKinect, consists of an infrared depth sensor and a custom software application; usability testing was carried out with fifteen therapists who performed two transfer assessments with the TransKinect. Proficiency in using features, usability, acceptability and satisfaction were analysed with validated surveys and themes were extracted from the qualitative feedback. Results The therapists were able to successfully complete the transfer quality assessments with 86.7 ± 5.4% proficiency. Total scores for System Usability Scale (77.6 ± 14.7%) and Questionnaire for User Interface Satisfaction (83.5 ± 8.7%) indicated that the system was usable and satisfactory. Qualitative feedback indicated that TransKinect was user-friendly, easy to learn, and had high potential. Discussion The results support TransKinect as a potential clinical decision support system for therapists for the comprehensive assessment of independent transfer technique. Future research is needed to investigate the utility and acceptance of TransKinect in real clinical environments. Implications for Rehabilitation Machine learning and computer vision can be used to analyze transfer technique TransKinect is a usable and user-friendly means for therapists to automate analysis Summary reports and videos of transfers show high potential for clinical use Adoption of TransKinect can increase quality of care for manual wheelchair users

Suppressing the HIFU interference in ultrasound guiding images with a diffusion-based deep learning model

Background and objectivesIn ultrasound guided high-intensity focused ultrasound (HIFU) surgery, it is necessary to transmit sound waves at different frequencies simultaneously using two transducers: one for the HIFU therapy and another for the ultrasound imaging guidance. In this specific setting, real-time monitoring of non-invasive surgery is challenging due to severe contamination of the ultrasound guiding images by strong acoustic interference from the HIFU sonication. MethodsThis paper proposed the use of a deep learning (DL) solution, specifically a diffusion implicit model, to suppress the HIFU interference. We considered the images contaminated with HIFU interference as low-resolution images, and those free from interference as high-resolution. While suppressing HIFU interference using the diffusion implicit (HIFU-Diff) model, the task was transformed into generating a high-resolution image through a series of forward diffusion steps and reverse sampling. A series of ex-vivo and in-vivo experiments, conducted under various parameters, were designed to validate the performance of the proposed network. ResultsQuantitative evaluation and statistical analysis demonstrated that the HIFU-Diff network achieved superior performance in reconstructing interference-free images under a variety of ex-vivo and in-vivo conditions, compared to the most commonly used notch filtering and the recent 1D FUS-Net deep learning network. The HIFU-Diff maintains high performance with ‘unseen’ datasets from separate experiments, and its superiority is more pronounced under strong HIFU interferences and in complex in-vivo situations. Furthermore, the reconstructed interference-free images can also be used for quantitative attenuation imaging, indicating that the network preserves acoustic characteristics of the ultrasound images. ConclusionsWith the proposed technique, HIFU therapy and the ultrasound imaging can be conducted simultaneously, allowing for real-time monitoring of the treatment process. This capability could significantly enhance the safety and efficacy of the non-invasive treatment across various clinical applications. To the best of our knowledge, this is the first diffusion-based model developed for HIFU interference suppression.

Innovating Healthcare Delivery: Enhancing Patient Engagement with Rule-Based Machine Learning Solutions

In contemporary healthcare, chatbots, driven by rule-based machine learning and sophisticated algorithms, are revolutionizing traditional models. This study explores their use in primary care triage, mental health support, chronic disease management, and telemedicine, highlighting their potential to enhance patient engagement, optimize clinical workflows, and improve care quality. Despite their promise, chatbots face challenges such as safeguarding patient privacy, mitigating biases, ensuring regulatory compliance, and maintaining ethical standards. The paper emphasizes the need for interoperability, scalability, and usability for seamless integration into healthcare systems. Future advancements include integrating advanced machine learning, evolving voice recognition, and personalized medicine approaches. By leveraging data analytics and predictive modelling, chatbots could proactively identify health risks, leading to a new era of preventive and personalized healthcare. Overall, this research highlights chatbots' profound impact on promoting patient-centred care, initiative-taking health management, and data-driven decision-making, fostering equitable access to high-quality healthcare services. Keywords— Rule-based chatbots, Machine Learning, Healthcare, Natural Language Processing, Patient Engagement, Clinical Support, Case Studies, Prospects.

Trends in forecasting groundwater ingresses into underground structures

Often, underground structures are faced with groundwater ingresses during their erection and even during their operation. To conceive the most suitable drainage or dewatering systems, and at the same time better guarantee the sustainability of these structures, these inflows should be accurately forecasted in advance. To this end, researchers have made considerable efforts and developed various solutions. This article put forwards the recent trends and progress related to the prediction of groundwater ingresses in underground structures. Pioneering solutions (analytical, semi-analytical, empirical and semi-empirical) as well as numerical, machine learning and other solutions are widely highlighted. Besides, the paper explains that the ideal solutions are still subject of current and future investigations. The need to continually opt for better schemes or strategies for accurate groundwater ingress prediction solutions is adequately expressed. Relevant inspirations can be drawn from this article for future accurate groundwater ingress forecasting solutions.

'Ripple effects' of urban environmental characteristics on cognitive performances in Eurasian red squirrels.

Urban areas are expanding exponentially, leading more species of wildlife living in urban environments. Urban environmental characteristics, such as human disturbance, induce stress for many wildlife and have been shown to affect some cognitive traits, such as innovative problem-solving performance. However, because different cognitive traits have common cognitive processes, it is possible that urban environmental characteristics may directly and indirectly affect related cognitive traits (the ripple effect hypothesis). We tested the ripple effect hypothesis in urban Eurasian red squirrels residing in 11 urban areas that had different urban environmental characteristics (direct human disturbance, indirect human disturbance, areas of green coverage and squirrel population size). These squirrels were innovators who had previously repeatedly solved a food extraction task (the original task). Here, we examined whether and how urban environmental characteristics would directly and indirectly influence performance in two related cognitive traits, generalisation and (long-term) memory. The generalisation task required the innovators to apply the learned successful solutions when solving a similar but novel problem. The memory task required them to recall the learned solution of the original task after an extended period of time. Some of the selected urban environmental characteristics directly influenced the task performance, both at the population level (site) and at individual levels. Urban environmental characteristics, such as increased direct and indirect human disturbance, decreased the proportion of success in solving the generalisation task or the memory task at the population (site) level. Increased direct human disturbance and less green coverage increased the solving efficiency at individual levels. We also found an indirect effect in one of the urban environmental characteristics, indirect human disturbance, in the generalisation task, but not the memory task. Such an effect was only seen at the individual level but not at the population level; indirect human disturbance decreased the first original latency, which then decreased the generalisation latency across successes. Our results partially support the ripple effect hypothesis, suggesting that urban environmental characteristics are stressors for squirrels and have a greater impact on shaping cognitive performance than previously shown. Together, these results provide a better understanding of cognitive traits that support wildlife in adapting to urban environments.

Deep reinforcement learning algorithms for dynamic pricing and inventory management of perishable products

A perishable product has a limited shelf life, and inefficient management often leads to waste. This paper focuses on dynamic pricing and inventory management strategies for perishable products. By implementing effective inventory control and the right pricing policy, it is possible to maximize expected revenue. However, the exponential growth of the problem size due to the shelf life of the products makes it impractical to use methods that guarantee optimal solutions, such as Dynamic Programming (DP). Therefore, approximate solution algorithms become necessary. We use Deep Reinforcement Learning (DRL) algorithms to address the dynamic pricing and ordering problem for perishable products, considering price and age-dependent stochastic demand. We investigate Deep Q Learning (DQL) solutions for discrete action spaces and Soft Actor-Critic (SAC) solutions for continuous action spaces. To mitigate the negative impact of the stochastic environment inherent in the problem, we propose two different DQL approaches. Our results show that the proposed DQL and SAC algorithms effectively address inventory control and dynamic pricing for perishable products, even when products of different ages are offered simultaneously. Compared to dynamic programming, our proposed DQL approaches achieve an average approximation of 95.5 % and 96.6 %, and reduce solution times by 71.5 % and 79.9 %, respectively, for the largest problem. In addition, the SAC algorithm achieves on average 4.6 % and 1.7 % better results and completes the task 56.1 % and 48.2 % faster than the proposed DQL algorithms.

Vocational school students’ use and opinions of voluntary-based online learning solutions presented in a mathematics course

This article focuses on the use made by Finnish vocational upper secondary students, including their opinions, with regard to the voluntary-based aspects and activities of a compulsory mathematics online course. In particular, the study investigated whether the vocational field affects students’ views and actions. There were altogether 313 students from six vocational fields participating in this study, but the number of students who responded to the feedback questionnaires related to the different opportunities and activities varied. Of the opportunities offered, the repetition opportunity was the most used and the most useful from the students’ point of view. Almost 90% of the respondents ended up, either voluntarily or under guidance, taking advantage of the repetition option. The education video was watched by almost 60% of the students responding to the questions related to the video. Students had a positive attitude towards the effects used in the educational video, the length of the video, and the level of explanation of theory. Approximately half of the respondents who were offered an additional practice opportunity took advantage of this, and the opportunity was seen as quite useful. There were only minor, if any, statistically significant differences between the vocational fields regarding the use of the opportunities offered and the opinions related to them. This research will assist in developing online teaching of mathematics at vocational upper secondary level and will also offer guidelines for the development of online education more generally.

Navigating the Evolution of Game-Based Educational Approaches in Secondary STEM Education: A Decade of Innovations and Challenges

The need to support science, technology, engineering, and mathematics (STEM) learning in secondary education is reflected in the ongoing investigation of innovative pedagogical practices, including game-based learning (GBL). Using an analysis of scholarly publications based on word co-occurrence, this study aimed to identify the main research themes addressed in the past decade by the scholarly community on game-based teaching and learning solutions in the context of STEM education in secondary schools, their evolution over time, and the key issues addressed in recent years. After a systematic selection, the titles and abstracts of the publications were collected in a text corpus and analyzed using T-LAB software version 7.2.1.4 (2022). A preliminary visual exploration of the keywords was performed to obtain an overall view of the issues addressed by the research. Specificity analysis was then applied to identify, for each subset of the corpus identified by the years of publication, the evolution of themes reflected in a change in the frequency of lemma use. Finally, to explore the most recent topics, the main thematic clusters of publications in the last three years were identified (thematic analysis of elementary contexts). The results suggest some changes in the issues addressed over the past decade, such as a shift in focus from the specific technologies and competitive elements of games to understanding how GBL can support engagement, motivation, and understanding of complex scientific concepts. The five key thematic clusters identified (“Experience”, “Application”, “Validation”, “Emotion”, and “Programming”) also indicate a stronger emphasis by the latest publications on the experiential and emotional components of learning, the need for empirical studies, and the integration of computational thinking and coding into GBL. Overall, this study indicates that GBL has the potential to become an integrated component of STEM education, evolving with pedagogical and technological innovations.

Robust Structure-Aware Graph-based Semi-Supervised Learning: Batch and Recursive Processing

Graph-based semi-supervised learning plays an important role in large scale image classification tasks. However, the problem becomes very challenging in the presence of noisy labels and outliers. Moreover, traditional robust semi-supervised learning solutions suffers from prohibitive computational burdens thus cannot be computed for streaming data. Motivated by that, we present a novel unified framework robust structure-aware semi-supervised learning called Unified RSSL (URSSL) for batch processing and recursive processing robust to both outliers and noisy labels. Particularly, URSSL applies joint semi-supervised dimensionality reduction with robust estimators and network sparse regularization simultaneously on the graph Laplacian matrix iteratively to preserve the intrinsic graph structure and ensure robustness to the compound noise. First, in order to relieve the influence from outliers, a novel semi-supervised robust dimensionality reduction is applied relying on robust estimators to suppress outliers. Meanwhile, to tackle noisy labels, the denoised graph similarity information is encoded into the network regularization. Moreover, by identifying strong relevance of dimensionality reduction and network regularization in the context of robust semi-supervised learning (RSSL), a two-step alternative optimization is derived to compute optimal solutions with guaranteed convergence. We further derive our framework to adapt to large scale semi-supervised learning particularly suitable for large scale image classification and demonstrate the model robustness under different adversarial attacks. For recursive processing, we rely on reparameterization to transform the formulation to unlock the challenging problem of robust streaming-based semi-supervised learning. Last but not least, we extend our solution into distributed solutions to resolve the challenging issue of distributed robust semi-supervised learning when images are captured by multiple cameras at different locations. Extensive experimental results demonstrate the promising performance of this framework when applied to multiple benchmark datasets with respect to state-of-the-art approaches for important applications in the areas of image classification and spam data analysis.

Playing with Leadership: A Multiple Case Study of Leadership Development Larps

Learning and development solutions often contain some form of role-playing activity. Many live-action role-playing (larp) leisure events openly center their themes or story on leadership topics like conflicting visions, cooperation, or strategic decision-making. Despite this, only a few larps to date have dared label themselves as “leadership development” edu-larps. This multiple-case study utilizes a synthesized theoretical framework that combines several typologies of both the leadership development research field and the fields of role-playing game studies, edu-larp, applied drama, and simulation. Four individual leadership development larps were identified, categorized by type, compared, and evaluated. The four cases have some striking similarities, as they all make use of the same learning cycle, have observers, and focus on developing versatility as a worthwhile leadership trait. Furthermore, all four larps have relatively few visible game mechanics, the hierarchical difference between characters is low and urgency or uncertainty is used to enhance learning. The main difference between cases is the setting, where a modern business setting seems to be the default, but more fantastical settings, such as “samurai in feudal Japan” were also used. Character design also differs, from abstract roles to very-detailed personality descriptions. Framing, such as duration, varies from one 2-hour session through five sessions spread over weeks to a continuous 36-hour session. These and other identified differences and similarities highlight patterns to build upon by designers and gaps to fill with later research.

LMA-EEGNet: A Lightweight Multi-Attention Network for Neonatal Seizure Detection Using EEG signals

Neonatal epilepsy is an early postnatal brain disorder, and automatic seizure detection is crucial for timely diagnosis and treatment to reduce potential brain damage. This work proposes a novel Lightweight Multi-Attention Network, LMA-EEGNet, for diagnosing neonatal epileptic seizures from multi-channel EEG signals employing dilated depthwise separable convolution (DDS Conv) for feature extraction and using pointwise convolution followed by global average pooling for classification. The proposed approach substantially reduces the model size, number of parameters, and computational complexity, which are crucial for real-time detection and clinical diagnosis of neonatal epileptic seizures. LMA-EEGNet integrates temporal and spectral features through distinct temporal and spectral branches. The temporal branch uses DDS Conv to extract temporal features, enhanced by a channel attention mechanism. The spectral branch utilizes similar convolutions alongside a spatial attention mechanism to highlight key frequency components. Outputs from both branches are merged and processed through a pointwise convolution layer and a global average pooling layer for efficient neonatal seizure detection. Experimental results show that our model, with only 2471 parameters and a size of 23 KB, achieves an accuracy of 95.71% and an AUC of 0.9862, demonstrating its potential for practical deployment. This study provides an effective deep learning solution for the early detection of neonatal epileptic seizures, improving diagnostic accuracy and timeliness.

The Efficiency and Effectiveness of 3D Printing Technology in Teaching Medicine and Surgery in a School of Medicine in the Developing World

Background: In the low and middle income countries’ healthcare system for medicine and surgery training and service provision, the use of computing technologies like 3D printing, would be of great help in creating faster, easier and transformational ways for training and doing surgical procedures, by creating software applications and hardware fabrications for efficient and effective learning solutions which are absent currently, hence leading to prompt treatment and surgical interventions in the developing world. Objective: The objective of this study was to examine the efficiency and effectiveness of the 3D printing technology in improving teaching and understanding of human anatomy and skills in the school of medicine in a developing country. Methodology: A randomized case-control study was used to analyze the impact of using 3D printed anatomical organ models as an adjunct to cadavers, in enhancing training of anatomy and skills development at Uzima University, School of medicine. Findings: There was substantial improvement amongst students taught with 3D printed organs together with cadavers in understanding the anatomy and skills taught as compared to those taught with cadavers alone. Students lectured with 3D printed brain and heart organ model took a median time of 2.2 hours IQR (1.6–2.8) for full identification of parts, while those taught with cadaver alone took a median time of 11.6 hours IQR (10.9-12.5), nearly 6 times more (p<0.001). In terms of narration of parts, functions and surgical procedures to treat ailments, those lectured with 3D printed brain and heart organ models took a median time of 6.4 hours IQR (5.6–7.2) for full narration, whereas those students lectured with cadaver alone took a median time of 17.5 hours IQR (18.8–18.8), p<0.001. Conclusion: We concluded that using anatomical models as an adjunct to cadavers improves training and skills of anatomy in medical schools in the developing world.

Multimodal deep learning for dementia classification using text and audio

Dementia is a progressive neurological disorder that affects the daily lives of older adults, impacting their verbal communication and cognitive function. Early diagnosis is important to enhance the lifespan and quality of life for affected individuals. Despite its importance, diagnosing dementia is a complex process. Automated machine learning solutions involving multiple types of data have the potential to improve the process of automated dementia screening. In this study, we build deep learning models to classify dementia cases from controls using the Pitt Cookie Theft dataset from DementiaBank, a database of short participant responses to the structured task of describing a picture of a cookie theft. We fine-tune Wav2vec and Word2vec baseline models to make binary predictions of dementia from audio recordings and text transcripts, respectively. We conduct experiments with four versions of the dataset: (1) the original data, (2) the data with short sentences removed, (3) text-based augmentation of the original data, and (4) text-based augmentation of the data with short sentences removed. Our results indicate that synonym-based text data augmentation generally enhances the performance of models that incorporate the text modality. Without data augmentation, models using the text modality achieve around 60% accuracy and 70% AUROC scores, and with data augmentation, the models achieve around 80% accuracy and 90% AUROC scores. We do not observe significant improvements in performance with the addition of audio or timestamp information into the model. We include a qualitative error analysis of the sentences that are misclassified under each study condition. This study provides preliminary insights into the effects of both text-based data augmentation and multimodal deep learning for automated dementia classification.

Metode Ummi Meningkatkan Kemampuan Membaca Al-Qur'an di Universitas Nahdlatul Ulama Al Ghazali (UNUGHA) Cilacap

This research aims to analyze the application of the Ummi method in improving the ability to read the Al-Qur'an at UNUGHA Cilacap. The aim is to provide a better understanding of the factors that influence the effectiveness of Al-Qur'an learning in the academic environment, as well as applying the ummi method to improve students' Al-Qur'an reading skills, with a focus on case studies at UNUGHA Cilacap. Through a case study approach, presenting an effective method for improving skills in reciting the Koran through the Ummi method. Reading the Koran proficiently is an important skill for Muslim students. However, many people face difficulties in mastering the correct recitation techniques. This research explores the effectiveness of the Ummi Method—an innovative teaching approach—in developing strong Al-Quran reading skills among students at Al Ghazali State Islamic University, Cilacap. Through the use of a qualitative case study design, we observed Ummi Method classes, interviewed enthusiastic students and dedicated tutors, and examined evaluation results. These findings reveal how the Ummi Method engages students with a fun yet rigorous step-by-step process. By utilizing repetition, an emphasis on clear pronunciation, and a caring learning environment, students gain confidence and improve their skills significantly. Real-life motivational examples from tutors also inspire spiritual and character development. Students appreciate learning at their own pace through this simple yet comprehensive system. The evaluation scores support that the Ummi Method improves the quality of reading, knowledge of Tajwid rules, and love of the Al-Quran. This research provides valuable insights into effective teaching approaches that address the needs of modern Muslim learners. The successful application of the Ummi Method at Al Ghazali State Islamic University Cilacap shows its potential in optimizing academic achievement and also strengthening faith in other educational contexts. This method is based on the rules of tajwid, makharijul letters, gharib, and tartil rules. Recommendations include curriculum improvements and facility support to improve students' ability to read the Koran. Further studies are needed for a deeper understanding. The findings of this research provide valuable insight for the development of effective learning strategies and solutions to overcome the challenges faced by students in studying the Koran.

Ensemble Random Forest-based Gradient Optimization based Energy Efficient Video Processing System for Smart Traffic Surveillance System

Deep learning solutions in big data applications can benefit cloud centres and can also lead to network communication overhead. Typically, data collected from traffic are sent to the traffic management centre for analysis. However, this process can worsen the network route to the traffic management centre. A two-tier mechanism has been developed to address this issue, which performs vehicle speed estimation and traffic congestion detection for efficient traffic management. The real-time traffic video data are captured and the video frames are initially processed through a foreground extraction process, which extracts the temporarily stopped vehicles on the road by removing background pixels from the frames. The video frames are then wrapped in an up-down view to remove the influence of the observation angle. The traffic congestion is then detected accurately based on the traffic characteristics using the proposed Ensemble Random Forest-based Gradient Optimization (ERF-GO) algorithm. The generalization error occurs when learning complex features on frames is minimized using a gradient-based optimization (GO) algorithm. Finally, the learned information on traffic conditions is forwarded to the cloud and edge computing environments based on network connection speed. The efficiency of the proposed ERF-GO is investigated in terms of performance metrics, namely root mean square error, speed detection error, execution time, computational cost, accuracy, latency, workload balance, precision, recall, f-measure, and congestion detection error rate. The analytic result displays that the proposed ERF-GO algorithm attains a greater accuracy rate of about 98.65% in detecting traffic congestion which is comparably higher than state-of-the-art methods.