Browse Items (3095 total)
Sort by:
-
Boosting Competitiveness Through Data: How Online Procurement Drives Data-Driven Decision-Making in Traditional Kirana Shops
Integrating traditional and modern elements presents significant challenges, yet when successful, the synergy can be immense. Onboarding Indian Kirana shopssmall, unorganized mom-and-pop storesinto a comprehensive digital infrastructure is crucial given the current retail landscape and evolving consumer demands. These shops are vital to the countrys food and grocery ecosystem but face disruption from the rise of e-commerce and organized retail. By adapting new business models and technologies, Kirana shops can enhance their competitiveness. This study highlights the critical role of digitalization and data-driven decision-making in small scale retail formats. Researchers collected primary data from Kirana shops doing online procurement and those relying on traditional methods like purchasing from distributors. Analysis of primary data shows that shops utilizing online procurement platforms demonstrate superior performance, attributed to factors like competitive pricing and timely delivery. Most importantly, the insights and analytics provided by eB2B platforms are game changers. Data emerges as the key differentiator; digitalization enables access to critical analytics, allowing for informed business decisions that improve success rates and provide a competitive edge. Consequently, this study propose an ideal digital end-to-end model designed to enhance operational efficiency and drive growth for unorganized Kirana shops. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Optimizing Car Recommendations: Power Analysis of Machine Learning Algorithms
The growing demand for efficient automobile recommendation systems has called for the need of algorithms that can proficiently assess and predict user preferences. This research focuses on the assessment of various machine learning algorithms, K-Nearest Neighbors (KNN), Decision Trees, Linear Regression, Weighted Scoring, and Content-Based Filtering. One of the main concerns of this study is to identify which recommendation algorithm is best suited for vehicle suggestions from an application perspective based on cost, mileage, engine size, fuel category, and user reviews. A dataset of 100 records was utilized to perform preliminary analyses so that algorithms were tested. Preprocessing procedures involved missing data handling, normalization of numerical features, and categorical variables encoding so that full precision predictions were obtained. Performances of algorithms were tested in terms of accuracy, scalability, and computational efficiency. Based on results, the highest accuracy was realized by Decision Trees with 85%, followed by Weighted Scoring at 82% and Linear Regression at 78%. Although KNN has an excellent accuracy of 74%, it is less scalable for very large datasets that are needed for an automobile recommendation system. The experimental results of this paper add to the evolving knowledge on the application of machine learning in the automobile world, again reinforcing the adequacy of Decision Trees as a valid technique for car recommendation systems. Recommendations for future studies include enhancing the database and exploring contemporary approaches to improve the accuracy of recommendations. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Machine Learning and Ensemble Models for Hazardous Asteroids Prediction
The prediction of hazardous asteroids near Earth is critical for planetary defense and avoiding any possible impacts. This study investigates the use of five ensemble models, XGBoost, Gradient Boost, CatBoost, Voting Classifier, and Random Forest, as well as four standalone machine learning models, K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Logistic Regression, and Decision Tree, to improve the prediction accuracy of identifying potentially hazardous asteroids. With 92% accuracy and 91% precision, Random Forest performed better than other models. It was the preferred choice for predicting hazardous asteroids because of its capacity to handle the hugedatasetwith efficiency and its ability tomanage non-linear data patterns. Additionally, XGBoost and CatBoost providedhigh accuracy at lowcomputational costs, making them suitable for real-time monitoring. KNN, on the other hand, did not perform well, and SVM's high processing time made it less useful. In particular, Random Forest ensemble modelperformed better at predicting hazardous asteroids. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
CMSFE: Cross-Model SSL Feature Extraction for Enhanced Remote Sensing Data Representation
Automatic Labeling of Remote Sensing Data fastens analysis in various applications such as environmental monitoring, urban planning, and disaster management. Supervised machine learning approaches rely on labeled datasets created through time-consuming processes. Creation of labeled datasets requires higher resources and such datasets are harder to obtain in most of the domains, and especially in Remote Sensing. This study proposes Cross-Model Self-Supervised Feature Extraction (CMSFE), a novel approach that enhances representation learning in unlabeled remote sensing datasets by integrating features from multiple pre-trained models and refining them through self-supervised learning (SSL). The extracted features are integrated to form a comprehensive and robust feature set that aids in separating different cluster of imagery. Experimental results with EuroSAT dataset demonstrate the quality of feature extraction in separating various classes without any manual intervention or labeling. Dimensionality Reduction and Manifold Learning is applied for visual interpretation of extracted feature space. These features can be further reused for analysis or modeling, highlighting the potential of SSL-based feature extraction methods in remote sensing to enhance representation learning and reduce dependency on labeled data. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Stride Insights: AI-Powered Field Position Forecasting System
This research discovers an AI-predictive model that uses a variety of machine learning algorithms to predict the top five finishers in a race. Horse racing is one of those paradigms that presents a challenging dataset against which the race outcomes can be predicted. Horse racing involves numerous variables: horse performance metrics, race conditions, and the jockey, together, decide the outcome of a race. To tackle such complexity, we test several algorithms, including CatBoost, Random Forest, k-Nearest Neighbors, Logistic Regression, Decision Trees, Support Vector Machines, Linear Regression, Naive Bayes, and Gradient Boosting, relating to the incorporation of categorical and continuous data. Our experiments demonstrate that the highest accuracy was achieved with CatBoost, which allows the model to handle categorical features well and is resistant to overfitting. The game theory component supplies useful elements in the strategic interaction between competing horses, thereby further increasing predictive accuracy. Performance metricsaccuracy, precision, and recall were used to estimate each model. The accuracy of CatBoost was found to be 74.1, while others were less accurate. This research provides an important resource for racing stakeholders, from trainers to punters. The research will be valuable in delineating race strategy and which horses are likely to win. This is an advancement in horse racing analytics and lays the foundation for predictive modeling to be explored in similar competitive environments in the future. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Data-Driven Malware Detection: Exploring Supervised Machine Learning Approaches
Malicious software must be detected in order to protect sensitive data and systems in the digital era, as sophisticated malware is posing serious risks to cybersecurity. By examining supervised machine learning approaches with a particular focus on Random Forest, Logistic Regression, and Decision Trees, this research proposes a data-driven approach to malware detection. These algorithms are trained to recognize patterns indicating malware by using labeled datasets containing four types of malwares, Ransomware, Trojan, Virus, and Worm. The performance of these algorithms is comprehensively investigated in the paper, with comparisons made between their accuracy, precision, recall, and F1-score. Based on the experimental results, Random Forest (96% accuracy) performed better in terms of robustness and accuracy of detection than both Logistic Regression (91%) and Decision Trees (84%). Logistic Regression provided faster computation at the expense of less accurate detection. Decision trees, while relatively simple to comprehend, performed moderately and they overfit the data. The studys conclusion highlights the significance of choosing the appropriate model in accordance with particular cyber security requirements, outlining the advantages and disadvantages of every approach as well as their practical applicability in real-time malware detection systems. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Dynamics of Socio-economic Factors in Shaping Fertility Rates in Manipur: A Robust Poisson Regression Framework Using NFHS-5 Data
This study utilizes Poisson regression to examine the total number of children ever born for all the surveyed women in Manipur, using data from the NFHS-5. Poisson regression is well-suited for modeling count data, such as fertility outcomes, due to its ability to handle discrete, non-negative variables. The model incorporates robust standard errors to deal with violations of the assumptions typically associated with Poisson models, such as under or over dispersion and heteroscedasticity in the residuals. Findings indicate that lower levels of education, age at first child, and limited health literacy are some of the significant predictors of higher birth rates. Furthermore, women from specific religious communities and economically disadvantaged backgrounds are likely to have more children, probably due to restricted access to family planning resources and healthcare services. These findings highlight the value for targeted policy interventions that prioritize improving womens educational opportunities and health literacy. Such strategies are essential for effectively managing fertility rates in Manipur and are in line with national sustainable development goals focused on health and education improvements. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Enhancing Data Security Through Semi-parametric Shrinkage Estimation of Shannon and Past Entropy in Geometric Distributions
The concept of entropy has been introduced in statistical methods to measure the amount of information contained in a random observation, and it plays a crucial role in various fields, especially in data security. This paper focuses on the semi-parametric shrinkage estimation of Shannon entropy and past entropy measures of the geometric distribution under complete, right, and time-censored sampling procedures. Shannon entropy, a key measure of uncertainty, along with past entropy (or min-entropy), which assesses the least predictable outcomes, plays a crucial role in ensuring strong data security, particularly in cryptographic systems and secure communications. While most existing literature addresses estimating these entropy measures for continuous distributions, this paper evaluates shrinkage estimators to enhance the efficiency of the ordinary semi-parametric least squares estimator for geometric distributions. This study explores the constant shrinkage factor and modified Thomson-type estimators, evaluating their effectiveness against traditional methods such as maximum likelihood estimators. Empirical investigations conducted with simulated samples indicate that shrinkage estimators consistently outperform maximum likelihood estimators, showcasing better relative efficiency. These results emphasize the potential of shrinkage estimators to enhance entropy-based measures in data security applications, which can lead to more robust cryptographic key generation, password strength analysis, and intrusion detection. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
A Deep Learning Approach to Clinical Decision Support in Heart Disease Diagnosis
Heart disease is the dominant cause of extinction worldwide, emphasizing the importance of early diagnosis and treatment planning. In this article, the authors developed a Clinical Decision Support System (CDSS) for heart disease prediction using deep learning techniques. This system will suggest a neural network architecture with Leaky ReLU as the activation function in the hidden layers and Sigmoid as the activation function in the output layer for binary classification. The configuration neural network is enhanced across three to nine hidden layers. The proposed approach is evaluated using accuracy as the measurable value on five multivariate datasets. By integrating advanced deep learning with clinical expertise, this study aims to enhance predictive accuracy, contributing to reduced heart disease mortality. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
AI and Big Data: Harnessing Data Science for Enhanced Consumer Insights
Data science is revolutionizing modern marketing strategies. These strategies enhance consumer intelligence and streamline decision-making processes. Data science techniques are helping businesses to get deep insights into consumer behavior, preferences, and emerging trends. In this paper, we focus on how the trident of artificial intelligence, machine learning, and quantum computing is reshaping marketing practices. Additionally, we focus on how artificial intelligence, machine learning, and quantum computing will impact data processing capabilities in future. The paper emphasizes the need for responsible data practices and discusses ethical issues such as data privacy and algorithmic bias. Several case studies focused on personalized marketing to improve customer satisfaction for companies like Netflix, Amazon, and Spotify are used. The findings suggest that businesses aiming to stay competitive will need to integrate data science in the complex data-driven world. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Estimating Function Approach for Modelling Non-Gaussian Time Series
In this study, an effective technique for estimating parameters of autoregressive models with non-Gaussian errors has been presented. For non-Gaussian error distributions, classical techniques like maximum likelihood estimation (MLE) are found to be computationally intractable. To address this issue, an estimating function (EF) technique has been utilized, which effectively estimates the model parameters by taking advantage of the error structure. Specifically, AR(1) with logistic errors has been used, and the optimal estimating functions have been derived by constructing martingale-based estimating equations tailored to logistic errors. A hybrid AR(1)-ANN model has also been developed to integrate the strengths of both linear AR(1) and nonlinear ANN models. The robustness and efficiency of the pro-posed approach in parameter estimation have been investigated using simulations, comparing its mean squared error (MSE) and bias to those of MLE. The applicability of this approach has been further demonstrated using both simulated and real datasets. The results show that, in the presence of non-Gaussian errors, the EF method provides a computationally efficient and reliable alternative to classical estimation methodologies. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Artificial Intelligence-Driven Perspectives on Maternal Health: Revealing Important Aspects and Improving Pregnancy Results via Machine Learning
A number of factors, including genetic, environmental and social ones, affect the intricate biological process of pregnancy. The developing foetuss health as well as the mothers must be maintained in the necessary secure equilibrium of these variables. The mothers health, which encompasses her mental as well as physical health, lifestyle decisions, money, social support systems and educational attainment, will determine whether the pregnancy ends well. Medical research has changed as a result of the long-awaited tools for processing for complicated datasets that have been made possible by recent advancements in machine learning models. These models have the ability to identify correlations between characteristics that are difficult for traditional analytical techniques to uncover. Therefore, scientists can improve their understanding of the elements influencing conception and create diagnostic tools by utilizing machine learning technology for timely intervention and customized treatment. Machine learning encompasses various techniques, such as logistic regression, linear regression, random forest, K-Nearest Neighbours and gradient boosting classifier. While Random Forest is an effective way to handle big databases with multiple dimensions and interactions, KNN classifiers are excellent for more organic, data-driven cluster finding of relevant instances and association investigation between various parameters and pregnancy outcomes. Logistic regression only explains the ways in which individual factors affect pregnancy outcomes; it cannot handle binary outcomes as well as linear regression does. We will look for significant determinants of pregnancy outcomes and assess each models performance. Important elements will also be expanded upon. Pregnant patients care, professional practice and improved program decisions may all benefit from this information. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Innovative Technology for Social Good: Real-Time Sign Language Generation Using TensorFlow
Real-time sign language is a basic means of communication for hearing-impaired people. There is a substantial communication barrier between sign language users and those who cannot comprehend sign language. Indian Sign Language (ISL) is built to ease the social challenge between hearing-impaired people and individuals who are unable to understand sign language using TensorFlow featuring Indian languages for smoother communication. The study aims to build a proficient real-time sign language translator using TensorFlow to detect hand signals in real-time video streams. Integration of TensorFlow enables real-time gesture detection, demonstrating how technology can bring about real progress when it comes to improving communication with persons who cannot hear. The application is trained on a specialized dataset comprising different Indian sign language signals, pre-processed to improve gesture recognition focusing on fast and accurate sign language recognition and translation. The objective is to develop a model that can recognize and translate hand gestures into text in Indian languages. This approach uses TensorFlow object detection API to recognize body gestures from real-time videos. The model is trained on a unique dataset of diverse Indian languages that is pre-processed for better recognition accuracy. Techniques like transfer learning are employed to fine-tune the model by integrating CNN for gesture recognition. The detected outputs are after-ward transformed into Indian languages. The systems accuracy may be restricted because of the quality and variety of different Indian languages across the country. The findings indicate that the model can accurately translate the collection of sign languages into text highlighting the potential of TensorFlow Object Detection for real-time sign language. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
BCI Radiology Images Converting into Report Using BERT and GPT
The construction of precise radiology reports from medical images is an essential aspect of Contemporary healthcare. Medical images such as X-rays, MRIs, CT scans, or ultrasounds. Also, it can make use of medical reports. Medical report has a bunch of details about each patients medical history, diagnosis, treatment plan, lab results, and more. This paper represents a theoretical examination. The paper mainly focuses on two prominent NLP models. One is BERT (Bidirectional Encoder Representations from Transformers) and the other one is GPT (Generative Pre-trained Transformer). This paper is going to validate their applicability to transforming brain-computer interfaces (BCI). This paper will utilize these radiology images in perfectly framed medical reports. By differentiating these models based on their Architectural properties, Linguistic processing abilities, and capability for clinical integration, this papers goal is to establish the most effective method for automated medical reporting. Merging of these insights from existing studies recommends that when BERT leads in context-based precision and getting an idea of complex medical terminology, GPT offers outstanding text-generation potential. This paper proposes that an intermixture procedure taking advantage of the strengths of both models may offer the most supreme solution for automated medical reporting, balancing precision with readability and clinical applicability. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Enhancing Reliability and Accuracy in Wind Energy Forecast Using CNN-LSTM Hybrid Model
The global shift to sustainable energy is increasing the demand for wind energy. Accurate forecasting becomes crucial for renewable energy systems to function effectively in terms of resource allocation, grid management, and overall reliability. The need for wind energy is growing as a result of the worlds transition to sustainable energy. For renewable energy systems to operate efficiently in terms of resource allocation, grid management, and overall reliability, accurate forecasting becomes essential. It is challenging for current forecasting methods to correctly predict the dynamic nature of wind energy demand. For utilities and grid managers, the inherent variability and unpredictability in wind energy generation pose serious issues. The forecasting models that are now in use are challenged by the variable and sporadic character of wind energy generation. This makes it more difficult to integrate wind energy into the electrical grid effectively and increases the risk of grid instability and inefficient resource utilization. This research addresses these challenges by proposing a hybrid forecasting model that integrates the strengths of Long Short-Term Memory (LSTM) networks and Convolutional Neural Networks (CNNs). By capturing both spatial and temporal dependencies in wind data, the hybrid model aims to enhance accuracy and reliability in wind energy forecasts. The precise forecasting of wind energy is made more difficult by shifting weather patterns, changing environmental factors, and shifting patterns of energy usage. Improving the forecasting models accuracy and dependability in the renewable energy industry requires addressing these difficulties. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Leaf Disease Analysis and Crop Suggestion Based on Soil Classification
Agriculture is a major contributor to the Indian financial system. As we realize approximately 60% of the population of Indiansrely on agriculture. Nowadays among the farmers who are educated, people also do agriculture. Farmers are dealing with problems related to low earningsbecause of loss of productivity. Farmers are thinking in the event that they use extra fertilizers they may get a precise yield, but it can grow the greater funding. If they do like this the bodily properties of soil may additionally decrease, and they are able to get the expected yield. To conquer this trouble, farmers have to realize which crop might healthy the specific piece of land. If they pick out the right sort of crop that is cultivated then robotically, the yield of the crop will grow. Hence, crop advice systems can be very beneficial to farmers. The yield of the crop can also depend on many factors like pH, nitrogen, phosphorus, potassium, and rainfall. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Stress Monitoring and Intervention for Women Using Enhanced BERT Models
Todays competitive world empowers women to lead their own lives, fostering innovation and inevitable growth in the industry. However, inclusivity in the corporate world is still a challenge for women. They have the pressure to meet societal expectations, familial responsibilities, and physiological changes associated with different stages of life. A concerning trend in society is the exponential increase in stress levels among women, a reflection of the evolving challenges and demands that women encounter to make themselves stand out from the rest. Various researchers have found that the number of women experiencing work-related stress is 50% higher than the number of men of the same age. The growing use of wearable Internet of Things (IoT) devices provides an opportunity to expand stress monitoring and intervention techniques in various situations, especially those that might be life-threatening. This study sought to confidently gain insights on stress monitoring for women through data collected by wearable IoT devices by segregating data based on device types. Applying a classification algorithm (transformer model) to determine the accuracy of stress indicators for these devices led us to build the stress accuracy prediction model. The existing BERT model is enhanced to process data beyond plain text. It is designed to uncover hidden patterns and trends associated with womens stress levels based on their pulse rates. These comparisons are portrayed using data visualizations. Using this enhanced BERT model adapted from the existing numerical algorithm, categorical data from IoT wearable devices is tested to accurately predict stress levels among women. This analysis demonstrates high predictive accuracy, with the earring IoT device achieving the highest accuracy of approximately 92%, indicating the effectiveness of the proposed model in stress monitoring across different wearable devices (earring, ring, and shoe). The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Privacy-Preserving Federated Learning: Foundations andAlgorithmic Directions
Federated Learning (FL) stands at the forefront of decentralized machine learning, revolutionizing collaborative model training among distributed devices while maintaining stringent privacy standards. FL requires multiple algorithms to handle issues with model initialization, synchronization, and convergence in remote environments. This paper comprehensively examines FL algorithms, focusing on pivotal techniques such as client-side training, server-side aggregation, and FedAvg. Detailed analysis elucidates these algorithms intricate workings, showcasing how they harmonize the aggregation of local model updates with global parameter refinement, thereby striking a delicate equilibrium between privacy preservation and model accuracy. The foundations of FL and the specifics of its sophisticated algorithms are covered in this study. By providing researchers with a roadmap for delving into FL algorithm development, this paper catalyzes unlocking new avenues of innovation and advancing the frontiers of privacy-preserving machine learning. For experimental learning, the federated learning implementation is carried out using the Flower framework on the well-known iris flower classification problem, with performance metrics thoroughly evaluated. Moreover, this paper represents, to our knowledge, the first work that extends the algorithmic directions presented in a review paper with detailed implementation on a sample problem, further encouraging exploration of various algorithms in FL implementation. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Smart Heritage: Leveraging Technology for Cultural Preservation and Sustainable Urban Development in Bangalore, Karnataka
In order to promote urban sustainability in Bangalore, this study explores how technology and data-driven solutions affect cultural tourism. In order to improve historical protection and handle urban issues with waste management, public services, energy efficiency, and transportation, the paper investigates artificial intelligence (AI) and smart tourist technologies. With Urbanization, commercialization, ecological deterioration, and changes in the quantitative and qualitative population dynamics, architecture is a risk and alarming situation for cultural formation of the city. This paper focuses on a qualitative study with case study on interventions for the cultural heritage of Bangalore. The study instrument utilized in the case of qualitative interviews makes aware of the historical nature of the skewness in the patterns of conservation and technology along with the stirred related issues through the historians, government authorities, and other influential personalities of the community. The study presents the possibilities of smart technologies in the field of digitalization and the use of virtual and augmented reality, as well concerns the use of environmental-friendly materials in the process of preservation. Specifically, the conclusions of the presented study concern the need to enhance the focused awareness and involvement of people, as well as the adequate governmental actions in relation to the issue of the preservation of birds. Therefore, the presented paper sets out a theoretical framework that defines the challenges, discussed strategies, and possible resulting impacts Thus, the papers goal entails advancing the policy-making process both on the national and local levels as well as the actions of proactive communities that would like to maintain Bangalores cultural identity for future generations to inherit. The notion of using integration technology in the management of stewardship of heritage also presents a useful means of managing past capitals for use in the future augmenting the vitality and sustainability of the citys economy. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025. -
Utilizing Deep Learning Features to Categorize WBCs in Blood Smear Images
Automated categorization of white blood cells (WBCs) is essential not just to identify infections, autoimmune ailments, and blood-related disorders, but also in the pivotal decision-making process concerning patient treatment and the efficient management of diseases. In this paper, an advanced approach for WBC type classification using smear images is proposed. The VGG16 model is utilized to capture intricate features of the images, which are then provided to an XGBoost classifier. This integration enables precise classification into 5 distinct WBC types. Our model shows a significant accuracy score of 92.3%, demonstrating its capability in accurately identifying WBC types from smear images. Proposed technique provides a promising pathway for automating WBC classification, thereby enhancing efficiency in disease diagnosis and decision-making within clinical settings. The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.