Browse Items (16481 total)
Sort by:
-
Machine Learning-Based Classical Dance Mudra Recognition Model
In this research, symbolic hand mudras of the Indian traditional dance style of Bharatanatyam are recognized and categorized using deep learning techniques. The three main goals are establishing baseline datasets to identify and categorize hasta mudras, designing an automated tutoring program for prospective students, and constructing a system for recommending videos that support cultural heritage. The research achieves a real-time recognition accuracy of 85% to 95% using convolutional neural networks (CNNs) and the Mobile Net architecture. This activity greatly aids virtual learning during pandemics, worldwide cultural relations, and preserving intangible cultural assets. The three main goals of this research are to establish baseline datasets for accurate mudra identification, create an automated tutoring program for participants, and build a video recommendation system to promote cultural heritage globally. The benchmark datasets that are used to train the models are made up of high-quality photos and videos of mudras that are taken and annotated under the direction of experts. While the video recommendation system supports attempts to preserve culture and advance education, the automated tutoring system provides participants with a comprehensive virtual learning environment and tailored feedback. To ensure the survival and continued appreciation of Bharatanatyam around the world, our endeavor substantially enhances virtual education, deep learning, and cultural preservation. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Prediction of Crime Hotspots Using Machine-Learning Techniques
Crime prediction is critical in improving police strategies and implementing measures for crime prevention and control. In recent years, machine learning has emerged as a critical way to predictive analytics in this domain. However, few studies have thoroughly compared various machine-learning algorithms for crime prediction. This study investigates the predicting capacities of various machine learning and ensemble approaches using historical public property crime data from a large city in India. Five ensemble models, Random Forest, AdaBoost, CatBoost, Gradient Boosting Machine (GBM) and eXtreme Gradient Boosting (XGBoost) and Four machine learning models, K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Nae Bayes and Decision Trees are used for crime predictive analysis in this study. The XGBoost model outperformed the other models tested, based primarily on historical crime data. XGBoost being an ensemble approachcombines multiple weak classifiers to create an effective classifier. Every weak learner concentrates on the faults made by the preceding ones, enabling the model to refine its predictions and fix errors repeatedly. When compared with other models used in the study, this resultedin higher accuracy. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Prediction of Grandiose Narcissism Using Machine Learning
The Gen-Zs have the tendency to exhibit a sense of self-importance and superiority excessively over social media. This study intends to predict Grandiose Narcissism based on Instagram usage and Fear of Missing Out (FoMO) among young adults. The study was conducted on a sample size of 300 young adults, recruited using convenient sampling, residing in the state of Assam, India. This study employed various machine learning models, including Logistic Regression, Random Forest, Support Vector Machine (SVM), Gradient Boosting, K-Nearest Neighbours (KNNs), and Gaussian Naive Bayes, to analyse the predictors of Grandiose Narcissism. Results showed that machine learning algorithms, especially KNN (90.7%) and Random Forest (88.70%) predicted Grandiose Narcissism accurately based onFoMO, Self-Esteem, PAUM. Additionally, Area Under Curve (AUC) in the range of 0.850.91 indicated that the variables in the data set are being discriminated in the context of specificity and sensitivity thoroughly. Significant influence of grandiose narcissism and FoMO on Instagram usage highlighted the role of social validation in enhancing online engagement. Future studies can include these algorithms to deduce patterns and develop real timebots to provide psychologically safe online environment. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Exploring the Intersection of Engineering and Health Leveraging 3D Solutions to Tackle Complex Medical Challenges
The medical imaging industry is on the brink of becoming as tangible as the world around us, blurring the lines between the virtual and the real. From compact gadgets to fully tangible anatomical models, medicine and technology have always been intertwined, but never as closely as they are today. This section explores immersive technologies in medicine and healthcare, focusing on virtual, augmented, and mixed reality, with a particular emphasis on the groundbreaking applications of 3D technology in medicine. This integration has ushered in a new era of medical advancements, promising significant progress in diagnosis, treatment, patient care, and healthcare overall. The adoption of 3D techniques, such as magnetic resonance imaging (MRI), computed tomography (CT), ultrasound, and more, enables doctors and clinicians to view anatomical structures as if they were real-life models. This enhances clarity, precision, and conciseness while reducing the risk of medical errors. The virtual world promised by 3D technology is also expected to improve communication skills between healthcare professionals and patients, especially for young medical students. 3D printing has seen significant development in recent years, playing a critical role in various applications, including medicine. This section will focus on three key advancements in 3D printing: its combination with the internet as a delivery vehicle, its integration with medical imaging, and its use in tissue engineering and regenerative medicine. We will also delve into three-dimensional display technologies, such as monoscopic 3D displays, stereoscopic 3D displays, and autostereoscopic displays. Technology's impact is not solely measured by its failures; sometimes, the small successes can save lives that might otherwise be lost. This piece will illustrate how technology could revolutionize medicine and reveal the potential we have yet to fully realize due to fear. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
A Univariate and Multivariate Time Series Analysis for the Prediction of Maize Production in India
This study examines the use of time series analysis for predicting maize production in India. The objective is to analyze the relationship between maize productions, domestic consumption, exports, and to forecast maize production using various time series models. The study employs cointegration techniques such as Johansen's test, Engle-Granger test, and Granger causality test to determine the long-term relationship between the variables. The findings exhibit that there is a bidirectional causal relationship between domestic consumption and export series and between domestic consumption and production series and that all three variables are co-integrated. To forecast maize production, the study employs both multivariate and univariate time series models. The multivariate models used are vector auto regressive and vector error correction models, while the univariate models used are ARIMA (auto regressive integrated moving averages), Holts exponential smoothing, NNAR (neural network auto regression), K-nearest neighbors (KNN), and LSTM (long short-term memory). The best forecast model is selected on the basis of a comparison of three evaluation metrics: mean absolute square error, mean absolute percentage error, and root mean square log error. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Investigating Risk Factors for Enhanced Portfolio Performance: An AI Approach for Indian Midcap Market Analysis
This research investigates the potential of machine learning (ML) for constructing portfolios that outperform human-based management, specifically focusing on the Indian midcap market. The study compares AI-based portfolio compositions, optimised using various risk measures, to the holdings of top midcap mutual funds. In this research, the top five midcap mutual funds sectoral distributions, portfolio compositions, and AI-generated portfolios are examined. According to the research, there is significant performance potential in the AI-generated portfolio, particularly when taking shorter investment horizons into account. Portfolios that maximise the Sharpe ratio produced the best returns throughout the course of the test period for four out of the six sectors, according to the research statistics. Additionally, in order to shed light on the effectiveness and possible advantages of our strategy, our study compares the suggested technique to existing investing strategies that concentrate on particular corporations as well as well-established market benchmarks. The research shows that, particularly when taking shorter investment horizons into account, the AI-generated portfolio has great performance potential. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Feature Fusion Classification for Emotional Intelligence Using Peripheral Signals
Real-time emotion identification is an innovation in the field of humancomputer interaction, which is an essential and challenging task. The existing studies methods for identifying emotions include face, audio, and physiological signals. The study aims to develop a model for emotion classification to identify and interpret human emotions through skin temperature, respiration, and plethysmography. The study also includes analyzing and interpreting emotional states through ensemble models. The classification is based on the frequency domain signal components extracted using the Fast Fourier Transform (FFT), such as amplitude and frequency, to identify emotional states. Ensemble-based machine learning algorithms such as XGBoost and LGBM achieved the highest accuracy in classifying various emotional states. The study involves unimodal and ensemble methods to analyze the signals. The comparative classification rate of unimodal results with ensemble shows that it is the highest at 85.99%, achieved for sad emotions by XGBoost. Fusing modules like respiration, skin temperature, and plethysmography maintains the accuracy level for all four emotions. The unimodal temperature has the highest accuracy of 86.1% for calm, whereas the fusion model has maintained accuracy for all the emotional states. The feature amplitude is the most promising feature for the classification method, which attains an average of 83.2% for XGBoost. The applications enhance user experiences and contribute valuable help in psychology, mental health care, and HumanComputer Interaction. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Optimizing Phishing Email Classification Through Scalable Feature Extraction Using MapReduce
A bag of features (BOF) may be made using either map reduction techniques or a combination of a thesaurus and domain knowledge. This research presents the BOFMR (Bag of Features using MapReduce) and BOFWT (Bag of Features with Weighted Terms) algorithms, a scalable and efficient technique for processing large email datasets and generating feature vectors based on pre-defined characteristics. The outcomes from using both BOFs on identical datasets are compared. The algorithm leverages the parallel processing capabilities of the MapReduce framework to handle the extensive data, ensuring performance and scalability. When creating a bag of words from a training dataset, the BOFMR technique is useful. The map-reduce technique will help to create a bag of features faster even in case of a larger chunk of data. In this experiment, as data size was limited, the performance of map reduce was not measured. In another BOFWT approach, the building of BOF with domain knowledge by using the word thesaurus was a challenge. The experimental result shows that the results of BOFWT are nearer to the output of BOFMR, and both algorithms show the highest accuracy among other methods. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Machine Learning Research Methods for Identifying Inaccurate Content
Social media, especially when disseminating news, is a valuable information resource. The paper presents methods for detecting fake news, comparing their effectiveness, identifying existing problems, and describes the vectors of further development of this research area. The paper begins with a description of the relevance of the Fake News problem, which clearly describes the negative impact of false news on all spheres of human life. The following is a description of methods for detecting false news, starting from the usual rules of text analysis and ending with complex ML algorithms. In this paper, a comparative analysis of detection methods is carried out, which is based on criteria of efficiency and accuracy. The author identifies the main problems of existing methods related to data quality, changing Fake News formats and the difficulties of automatically determining the reliability of information. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Efficient Intrusion Detection through Class Balancing and Feature Selection: A Case Study with SVM
Intrusion Detection Systems are of paramount importance in network security. However, in real-world scenarios, they always suffer from the challenge of class imbalance, which is dominated by normal traffic. This paper presents a novel approach to enhancing the performance of IDS by proposing a hybrid of the Random Under sampling technique with the univariate feature selection technique, SelectKBest, for handling both problems of class imbalance and high dimensionality. This model was hence tried on the Bot-IoT dataset, which is a real-world IoT network traffic representation. The SVM classifier, which has been trained with the resampled and feature-selected data, showcased 95% balanced accuracy for both normal and malicious traffic detection. The combination of RUS and SelectKBest, apart from reducing overfitting, ensured the retention of the most relevant features and thereby made the IDS model robust. It can practically enhance the performance of IDS in an imbalanced and high-dimensional dataset by providing a balanced, efficient, and precise detecting mechanism. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
An Improved AI-Based Low Latency Data Transmission in 5G Communication Systems
This paper devised an advanced artificial intelligence (AI) solution for ultra-low latency data transmission in 5G networks. With increasing data rates and lower latency required in 5G networks, efficient methods for transmitting the maximum amount of data are necessary. We have developed an approach that uses AI algorithms so that data transmission can be done more optimally and help reduce latency, providing better overall performance. Our approach consists of several steps, in which we predict the traffic patterns using machine learning techniques in step 1 and allocate network resources accordingly. That helps reduce network congestion and speeds up data transmission. We also introduce deep learning algorithms to adjust the transmission parameters according to network conditions, reducing latency. We simulate our algorithm in 5G network scenarios to assess its performance. The comparison of the results shows that a very low latency was achieved for this design over the earlier methods. Our developed AI-based improved solution provides a potential key to low latency data transmission in 5G communication systems. Integrating AI methods makes the system not only perform better but also be able to adapt more easily when network conditions change. The next steps are to explore the improvements of algorithms and implement them practically in 5G networks. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Facial Emotion Recognition Augmented with CNNs and Face Detection: Toward Emotive Emoji Synthesis
Emotion recognition is a crucial component with broad applications in technology and healthcare industries specifically in humancomputer interaction. To improve emotion recognition accuracy, this research introduces an innovative technique that integrates face detection with Convolutional Neural Networks (CNNs). Using the Fer2013 dataset, the approach consists of carefully identifying faces in images as a preprocessing step, followed by training a CNN network to identify emotions and create corresponding emojis. After conducting extensive testing and assessment, it is determined that after employing a face detection algorithm the suggested framework is effective in both correctly identifying emotions and producing visually appealing emojis. This helps to create an interface for emotional communication that is more user-friendly and captivating. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Organizational Factors Impacting the Retail Industry's Adoption of Blockchain Technology
Indian retail is undergoing a transformation due to rapid digitization and shifting consumer preferences. Blockchain technology is changing retail chain management by boosting efficiency, security, and transparency. This technology can also alter Indian retail operations by enabling verified transactions and improving inventory management to build consumer trust. Blockchain technology adoption in retail depends on organizational readiness, technological knowledge, and top management support. This study examines organizational aspects affecting blockchain adoption in retail and develops and validates a model for organizational characteristics affecting blockchain adoption. Thus, the study examines three key blockchain adoption intention constructs of blockchain knowledge, organizational readiness, and organizational support. Retailers received surveys online and the data was analyzed using SEM. The study supports that organizational management support (p?=?0.017) and organizational readiness (p?=?0.008) are significant precursors to the intention to adopt this technology. The study concludes that organizations must support and improve their readiness for modern technology with the top management's cooperation. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Privacy-Integrity Aware Efficient Workflow Scheduler for Edge-Cloud Platform
Scientific workflow execution in edge-cloud platform have attained in-creased attention, due to reduction in overall makespan assuring workflow deadlines. The workflow task is composed of diverse subtasks which are either executed in edge and the cloud; they are prone to security risk. Any loss of security breach will result to privacy and data integrity issues. Thus, providing security and meeting workflow execution strict deadlines becomes extremely difficult. The current workflow scheduling methods failed to assure both privacy and integrity together under edge-cloud computing platform. In addressing the research security and efficiency issues, this article introduced a novel approach namely Privacy-Integrity Aware Efficient Workflow Scheduler (PIAEWS) for edge-cloud platform. The PIAEWS introduces a novel trust metrics to assure only authenticated node takes part communication and consensus model to assure data integrity without compromising on user privacy constraint. The PIAEWS improves makespan and reduces overall energy consumption by assuring both security and performance together when executing genome sequencing workflow application in edge-cloud platform. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Smart People Counting System by Enhancing Accuracy and Affordability with YOLOv5 and Cloud-Based Integration
Considering that all moving objects are humans, much of the work in data is based on recognizing and tracking moving objects. In this work, we present a method for counting peoples faces. Even though we use the face mask, the deep learning-based YOLOv5 algorithm and Faster R-CNN allow us to recognize the face. We do a very good job of counting people. To make the calculation more accurate, we introduced a new type of intelligent small scale computing system consisting of cheaper hardware and user-friendly cloud computing software. These findings show that intelligent computing systems can realize human vision. Additionally, by combining inexpensive hardware with cloud-based software, the planning process becomes more transparent and cost-effective. Finally, the web application allows users to view the number of authorized and unauthorized users. Based on the results obtained from this method, the deep learning YOLOv5 algorithm is used to identify and match human images to increase security, and thanks to cloud storage, users can easily view all calculated results, increasing the accuracy by 98.53%. Owing to the truth that most of the secure watches cannot be able to check each and each individual The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Detecting and Countering Misinformation Through NLP-Based Approach for Fake News Detection
The rapid expansion of digital media and the seamless transmission of information have raised serious concerns about the widespread dissemination of misinformation and fake news. Combatting this issue requires robust and effective techniques that can accurately detect and classify fake news. Natural language processing (NLP) approaches have emerged as powerful tools in this endeavor, leveraging advanced text classification algorithms to identify and counteract misinformation. This study includes NLP approaches for countering misinformation through text classification, with a specific focus on fake news detection. Leveraging natural language processing techniques, the project implements a text classification pipeline for identifying and distinguishing between genuine and fake news. The pipeline encompasses essential NLP steps such as tokenization and stop word removal. Traditional machine learning algorithms, such as the gradient boosting classifier, CatBoost classifier, random forest classifier, AdaBoost classifier, logistic regression, and SVM linear kernel are trained using the transformed data to classify news articles. This study explores feature engineering techniques and model evaluation to enhance the classification performance. Experimental results indicate the effectiveness of The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Reinforcement Learning for Early Detection and Intervention of Sepsis with Graph-Based Personalized Treatment Recommendations
Early detection and treatment of sepsis, a condition that can become fatal through the bodys response to infection, can enhance patient life. This paper explores how reinforcement learning can be applied to the early detection and treatment of sepsis, along with its novel features, which include personalized treatment recommendations and graph-based representations using Graph Neural Networks (GNNs). Moreover, domain adaptation and transfer learning strategies make the model applicable in a wide range of clinical contexts. The RL model is therefore designed to identify early warning signs and give prompt, individualized answers to avoid major repercussions. To ensure wide application, the RL model was trained using an enormous dataset of patient vitals, lab results, and clinical notes from numerous centers. It is already proven in real-life clinical situations that this model can improve patient outcomes and the quality of clinical decisions. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Reinforcement Learning for Early Detection and Intervention of Sepsis with Graph-Based Personalized Treatment Recommendations
Early detection and treatment of sepsis, a condition that can become fatal through the bodys response to infection, can enhance patient life. This paper explores how reinforcement learning can be applied to the early detection and treatment of sepsis, along with its novel features, which include personalized treatment recommendations and graph-based representations using Graph Neural Networks (GNNs). Moreover, domain adaptation and transfer learning strategies make the model applicable in a wide range of clinical contexts. The RL model is therefore designed to identify early warning signs and give prompt, individualized answers to avoid major repercussions. To ensure wide application, the RL model was trained using an enormous dataset of patient vitals, lab results, and clinical notes from numerous centers. It is already proven in real-life clinical situations that this model can improve patient outcomes and the quality of clinical decisions. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
ResFruitGrader: Leveraging Residual Networks for Advanced Fruit Quality Grading Systems
The rising agricultural industrys requirement for effective sorting and grading procedures has increased the demand for automated and precise fruit quality assessment in recent years. This study aims to attain high classification accuracy by investigating the use of Convolutional Neural Networks for fruit quality identification. As customers place a higher value on fresh and wholesome options, the agriculture and food industries must meet rising demands for premium produce. Fruit quality must be guaranteed since it directly affects consumer happiness and the profitability of the sector. Preprocessing methods, CNN model creation, training, and evaluation utilizing cutting-edge deep learning techniques comprise the methodology applied in our study. The research demonstrates the CNN-based methods stability and dependability in identifying a range of quality attributes, such as fruit imperfections, size, color, and maturity. The suggested CNN architecture performs remarkably well, recognizing fruit quality parameters with a 99.5% accuracy rate by utilizing a collection of various fruit photos. A promising path for improving efficiency and accuracy in fruit quality assessment within the agricultural industry is provided by the researchs insights into the transferability and scalability of the developed model for practical applications in automated fruit sorting systems. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
AIs Role in Semantic Segmentation for Data-Driven 3D Models of Heritage Structures
Using point cloud data from laser scanning and photogrammetry to create three-dimensional models with scan-to-BIM processes has become increasingly common in heritage conservation. During the processing of point clouds, semantically segmenting data can translate captured spatial information into intelligent data structures, enabling classified, accurate, data-driven digital models of heritage structures. Subsequently, digital models are utilized for analytical tasks like structural tests, energy optimization, etc. Artificial Intelligence (AI) has become a promising solution for automating Three-Dimensional Point Cloud Semantic Segmentation (3DPCSS), enabling a faster and more accurate composition of parametric objects within 3D modeling and management systems. However, implementing 3DPCSS solely with AI presents various technical and theoretical challenges. The geometrical complexities inherent in historical structures often necessitate manual segmentation processes or oversimplified representations that miss the unique characteristics of the building. Therefore, selecting an appropriate AI framework for 3DPCSS is essential to ensure accurate results. Multiple factors determine algorithms selection, making it challenging to categorize universal solutions. The paper highlights the key factors: 1) Data collecting tools and technologies, 2) Types of the dataset, 3) Complexity of geometrical elements, and 4) Computational tasks. AI frameworks are typically selected based on the suitability and significance of these factors relative to the projects intent. Very few studies talk about the choices of algorithms. This papers significant contribution is recognizing trends in effective data acquisition strategies through a case study in India. Additionally, it identifies state-of-the-art AI models from the past decade based on a systematic literature study. The paper infers the extensive use and advancement of hybrid approaches tailored to multi-modal data types and the specific needs of heritage projects. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.