Volume 14 | Issue 6 |

 Abstract

The alarming escalation of crimes against women worldwide necessitates urgent technological interventions capable of providing real-time safety mechanisms. This paper presents EmpowHer (also referred to as SafeHer), a production-grade women safety mobile application built on React Native with TypeScript. We analyse the system from a Senior Research Developer perspective, formally specifying both functional and non-functional requirements, proposing a layered microservices-oriented architecture, and justifying the complete technology stack. The design encompasses an Emergency SOS subsystem with sub-2-second trigger latency, continuous GPS tracking at five-second intervals, AES-256 encrypted local storage, trusted-contact management, offline SMS fallback, automated audio/video evidence recording, and an AI-driven risk-detection module.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Women Safety Application, React Native, Emergency SOS, GPS Tracking, AES-256 Encryption, Microservices Architecture, Real-Time Location Sharing, Mobile Security, Firebase Cloud Messaging, AI Risk Detection.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Now a day, World Wide Web has become a popular medium to search information, business, trading and so on. Various organizations and companies are also employing the web in order to introduce their products or services around the world. Therefore E-commerce or electronic commerce is formed. E-commerce is any type of business or commercial transaction that involves the transfer of information across the internet. In this situation a huge amount of information is generated and stored in the web services. This information overhead leads to difficulty in finding relevant and useful knowledge, therefore web mining is used as a tool to discover and extract the knowledge from the web. Besides, the security issues are the most precious problems in every electronic commercial process. This massive increase in the uptake of e-commerce has led to a new generation of associated security threats. In this paper we use techniques for security purposes, in detecting, preventing and predicting cyber-attacks on virtual space

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Ecommerce, Cybercrime, threats, security ,attacks.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This study presents a scalable and efficient hybrid recommendation framework designed to address the challenges of modern e-commerce systems, including data sparsity, cold-start problems, and real-time processing requirements. The proposed system integrates multiple recommendation strategies, namely content-based filtering, collaborative filtering, and deep learning-based approaches, to leverage their complementary strengths. User interaction data, such as clicks, views, and ratings, is collected and processed through a structured pipeline involving data storage, preprocessing, and feature engineering. The framework employs content-based techniques for feature extraction, collaborative filtering for user-item similarity modeling, and advanced methods such as matrix factorization and neural networks, including recurrent and transformer-based architectures, to capture complex interaction patterns. A hybrid filtering mechanism combines these approaches to generate highly personalized and context-aware recommendations. The system is further evaluated using standard performance metrics such as precision, recall, and RMSE, along with A/B testing to validate its effectiveness in real-world scenarios. Additionally, the proposed architecture supports scalable deployment through API-based services and enables real-time recommendation generation. A continuous feedback loop is incorporated to refine model performance based on user interactions. Experimental observations indicate that the hybrid approach improves recommendation accuracy, diversity, and robustness compared to traditional baseline models. Overall, the framework enhances user engagement and satisfaction while providing a practical solution for large-scale recommendation systems in competitive e-commerce environments.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Hybrid Recommendation, Collaborative Filtering, Content-Based Filtering, Deep Learning, Neural Collaborative Filtering, Matrix Factorization, E-commerce

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Exoplanet research has grown into one of the most compelling frontiers in modern astrophysics reshaping what we know about worlds beyond our own solar system. These are planets orbiting stars other than the Sun and detecting them is rarely straightforward, astronomers typically rely on indirect methods, with transit photometry being among the most productive [1]. Space missions like the Kepler Space Telescope and TESS have generated enormous quantities of observational data, collectively responsible for confirming thousands of exoplanet candidates [3][4].Even so, a large share of that data has historically required manual review, which is both slow and inconsistent in practice. Classical detection tools like the Box Least Squares algorithm do a reasonable job of flagging periodic dips in light curves, but they struggle with noise and frequently flag non-exoplanets as real candidates. This paper introduces a hybrid detection framework that draws on traditional algorithms alongside machine learning and deep learning methods. Combining Random Forest classifiers, convolutional neural networks and oversampling via SMOTE, the system targets three key improvements: fewer false positives, stronger detection accuracy and the ability to scale across much larger observational archives.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

exoplanet detection, machine learning, deep learning, light curve analysis, space exploration, transit photometry.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Precision agriculture requires real-time monitoring and data-driven decision-making to optimize crop yield and resource utilization. This paper proposes a smart soil analysis and crop prediction system that integrates IoT sensors, a Raspberry Pi microcontroller, and machine learning algorithms. The system continuously monitors soil moisture, temperature, and pH, stores data in a cloud database, and predicts the most suitable crops using trained ML models. A relay-controlled water pump enables automated irrigation, improving water-use efficiency and reducing manual intervention. Experimental results demonstrate a crop prediction accuracy of 94% compared to conventional methods. The proposed solution offers a scalable, real-time, and sustainable framework for enhancing agricultural productivity and efficient resource management.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

IoT, Smart Agriculture, Crop Prediction, Soil Monitoring, Machine Learning, Raspberry Pi, Automated Irrigation, Precision Farming.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Programming languages are generally designed using English-based syntax, which can make learning difficult for individuals who are more comfortable with regional languages. This paper presents Bhaashankit, a Marathi-based programming interpreter that allows users to write and execute programs using familiar linguistic constructs.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Marathi Programming, Interpreter Design, Regional Language Computing, Lexer, Parser, Abstract Syntax Tree, Programming Education

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This research presents an interactive platform for AI-driven artwork generation, transforming written text into expressive, high-quality visuals for users of all artistic skill levels. The system integrates a pretrained Stable Diffusion model with a robust MERN stack ( MongoDB,Express.js, React.js, Node.js) web application. This architecture provides an accessible digital canvas that instantly turns ideas into visuals, making it a powerful tool for rapid ideation, storytelling, and accelerating creative workflows for professional designers and artists. Key features include support for high-resolution output and a categorized style selection, such as Meme, 3D Cartoon, 3D illustration, Sketch, Ghibli, Abstract and Logo generation enabling precise artistic control. Ultimately, the system enhances productivity while preserving the uniqueness of each user's vision, offering a vital bridge between imagination and innovation.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Style Selection, Stable Diffusion, Generative AI, MERN stack, Artistic skills, high quality visuals

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Renewable Energy, Machine Learning, Energy Prediction, Optimization, Sustainability, Smart Grid

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Customer segmentation is a cornerstone of data-driven marketing, personalization, and decision-making. With the rapid growth of digital platforms, organizations now handle large-scale, high-dimensional customer data, making traditional segmentation techniques insufficient. This analytical study examines major clustering algorithms used for customer segmentation, evaluates their suitability for large datasets, and compares their performance in terms of scalability, accuracy, interpretability, and computational complexity. The proposed framework leverages Mini-Batch K-Means, a batch-based learning algorithm that processes data incrementally in small subsets, significantly reducing memory usage and improving execution speed. Experimental results demonstrate that the proposed scalable clustering framework achieves faster convergence and better resource utilization compared to conventional techniques, proving suitable for real-world big data applications.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Customer Segmentation, Clustering Algorithms, Mini-Batch K-Means, K-Means, DBSCAN, Scalable Clustering, Big Data, Unsupervised Learning

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Air Quality Index (AQI), Air Pollution, Regression Techniques, Machine Learning, PM2.5, PM10, Support Vector Regression, Random Forest Regression, Deep Learning, LSTM

  License

Creative Commons Attribution 4.0 and The Open Definition