Air quality is a critical concern in urban environments, impacting public health and quality of life. In order to keep track of the trend of air quality, it is especially important to forecast it accurately. The Air Quality Index (AQI) is used for reporting daily air quality. It tells you how clean or polluted your air is, and what associated health effects might be a concern for you. The daily average value of the mass concentration of PM2.5, PM10,SO2, NO2, CO, O3 was obtained for this study.By integrating meteorological factors, pollutant concentrations, and historical AQI
trends. This model offers a comprehensive and accurate forecast of AQI levels. The model not only improves the accuracy of short-term AQI predictions but also enhances our understanding of the complex interactions between various environmental factors affecting air quality, ultimately contributing to better-informed decision-making for pollution control and public health management. Air Quality Index (AQI) is the crucial foundation for measuring air quality, which reflects the influence of air quality on people’s health and life to a certain extent.
Spam emails continue to pose significant challenges to users and organizations, necessitating robust classification techniques for effective filtering. This paper presents an overview of spam classification methods leveraging unsupervised learning algorithms. By eschewing the need for labeled data, unsupervised learning offers a promising avenue for identifying and categorizing spam without explicit human intervention. The abstract discusses the key concepts and methodologies employed in unsupervised spam classification, encompassing clustering algorithms such as K-means, hierarchical clustering, and density-based methods.
Cyberattacks and the use of malware are more and more omnipresent nowadays. Targets are as varied as states or publicly traded companies. Malware analysis has become a very important activity in the management of computer security incidents. Organizations are often faced with suspicious files captured through their antiviral and security monitoring systems, or during forensics analysis. Most solutions funnel out suspicious files through multiple tactics correlating static and dynamic techniques in order to detect malware. However, these mechanisms have many practical limitations giving rise to a new research track. The aim of this paper is to tackle the use of machine learning algorithms to analyse malware and expose how data science is used to detect malware. Training systems to find attacks allows to develop better protection tools, capable of detecting unprecedented campaigns. This study reveals that many models can be employed to evaluate their detectability. Our demonstration results illustrate the possibility to analyze malware leveraging several machine learning (ML) algorithms comparing them.
This project aims to generate word embeddings for Sanskrit shlokas by employing a hybrid approach that combines traditional linguistic knowledge with modern computational techniques. Sanskrit, an ancient Indo-Aryan language, holds significant cultural and religious importance, with shlokas serving as fundamental units of expression in various texts. The proposed methodology involves collecting a diverse dataset of Sanskrit shlokas and preprocessing them through tokenization, sandhi splitting, and normalization. Traditional linguistic features, including morphological, phonological, and semantic aspects, are extracted from the shlokas. These features are then integrated with modern natural language processing (NLP) techniques. The project aims to contribute to the preservation and understanding of Sanskrit language and literature while also showcasing the potential of hybrid approaches in computational linguistics.
Platforms for social media like Facebook, Twitter, Instagram, and others have a big impact on our lives. Our lives nowadays rely heavily on social media. Everyone uses social media, whether it be to share beautiful, expensive photos, follow celebrities, or talk with nearby and distant pals. It is a fantastic place for exchanging knowledge and interacting with others. However, everything has a drawback. Social media has a significant role in our lives, yet there have been times when it has become problematic. People are actively participating in it the world over. However, it also has to deal with the issue of bogus profiles. False accounts are frequently created by humans, bots, software, or machines. They are employed in the spread of rumors and illegal actions like phishing and identity theft. The aim of this project is to use several machine learning techniques to discriminate between fake and authentic profiles. It is possible to disable or delete a fake profile when it is found, preventing cyber security issues.
This project focuses on automating the detection of middle ear pathologies using the VGG16 convolutional neural network. Trained on a diverse dataset of annotated images from various imaging modalities, including otoscopy, CT, and MRI, the fine-tuned VGG16 model shows promising results. Its accuracy in identifying and classifying middle ear disorders is evaluated against traditional methods, showcasing advantages such as rapid image analysis and potential integration into healthcare systems. The VGG16-based approach holds promise for improving diagnostic accuracy, enabling early intervention, and enhancing patient outcomes in managing middle ear disorders. Further research is needed to validate its performance on larger datasets and explore integration into clinical practice.
This project introduces an innovative approach to plant disease detection using machine learning, specifically convolutional neural networks (CNNs). By analyzing digital images of plant leaves and incorporating environmental factors, the project can accurately detect diseases early on, promoting sustainable agriculture. The machine learning models are trained on a diverse dataset, ensuring adaptability across different crops and diseases. The project’s user-friendly interface allows farmers to receive real-time feedback, empowering them with actionable insights for effective crop management. Overall, this automated plant disease detection project aims to enhance crop productivity, reduce losses, and strengthen agricultural resilience.
Explore an intelligent and adaptive virtual assistant implemented in Python, revolutionizing human-computer interactions. Leveraging natural language processing (NLP) and machine learning, this system excels in voice commands, chat interactions, and dynamic responses tailored to user preferences. Adapting based on user history, it integrates external APIs for diverse functionalities and ensures scalability for additional features. Evaluation metrics demonstrate its effectiveness in accurate interpretation, relevant information provision, and dynamic conversational adaptability. This research contributes to human-computer interaction, presenting a framework for intelligent virtual assistants using Python. With applications in smart home automation, personal productivity, and accessibility tools, it highlights Python’s significance in creating adaptive virtual assistants.
This Python implementation introduces a car sound-based classification system utilizing Convolutional Neural Networks (CNN), implemented with TensorFlow and Keras. Trained on a diverse dataset, the CNN effectively distinguishes car-related sounds, displaying high accuracy. The system’s robustness to environmental variations positions it for applications in automotive diagnostics, smart cities, and intelligent transportation. This research contributes to audio signal processing and machine learning, providing a scalable solution for categorizing car sounds across diverse settings, fostering advancements in automotive technologies.
Exploring the realm of biometric authentication, this study delves into dorsal hand vein patterns as a distinctive and secure means of identity verification. Employing non-intrusive near-infrared sensors, dorsal hand vein images undergo sophisticated processing using advanced machine learning algorithms, including Convolutional Neural Networks (CNNs). The resulting authentication system undergoes rigorous training, validation, and evaluation using key performance metrics. Comparative analyses underscore the unique advantages of dorsal hand vein authentication, while addressing practical considerations such as user acceptability and ethical implications. This non-intrusive and forgery-resistant biometric method contributes significantly to the field of secure identification, finding potential applications in finance, healthcare, and secure facility access.
