Volume 12 | Issue 5

 Abstract

This research study presents a novel approach that integrates state-of-the-art Convolutional Neural Networks, algorithms to transform the management of women's health. This project intends to employ artificial intelligence to build a comprehensive manual that is unique to the health requirements of women, given the increasing prevalence of modern technology in healthcare. The suggested method covers a broad range of health-related topics, such as illness prevention, mental health, dietary counseling, and reproductive health, among others. CNN algorithms are used to provide real-time insights and recommendations based on lifestyle choices, medical history, and demographic details. This allows the guide to adapt to each user's individual health profile. The freedom of women to take proactive measures to optimize their health outcomes is at the heart of this research. Through intuitive design, the resource guide promotes self-care habits and educated decision-making. By giving users access to peer networks, professional guidance, and related information, it also helps users feel more supportive of one another and connected to each other.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

CNN algorithm, personalized healthcare, disease prevention, wellness management, artificial intelligence, women's health, and community support

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Phishing, a prevalent cybercrime, involves deceiving individuals into disclosing personal or confidential information under the guise of legitimate websites or emails. Recognizing phishing websites is challenging due to their similarity to authentic ones. Our study focuses on employing machine learning techniques for efficient phishing website detection. We detail the approach encompassing data gathering, preprocessing steps, extracting features, and employing Support Vector Machine (SVM) for classification. SVM stands out due to its capacity to handle high-dimensional data and non-linear relationships, making it robust to overfitting. SVM offers understandable detection capabilities by optimizing the margin between classes to its maximum extent. Our research aligns with broader cybersecurity objectives, aiming to safeguard individuals and organizations against online deception. Through the development of robust detection systems, we contribute to enhancing cybersecurity measures and empowering users with more secure online experiences. This endeavor underscores the importance of proactive measures in combatting evolving cyber threats

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Phishing, Support Vector Machine, high- dimensional, Feature extraction

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Social media platforms like Twitter face a growing challenge with the proliferation of malicious Twitter bots. These bots spread disinformation, manipulate public opinion, and engage in fraudulent activities, undermining trust in online spaces. To address this issue, our project focuses on detecting malicious Twitter bots using advanced machine learning techniques and URL analysis. By examining various features, including URL characteristics, we've developed a robust framework. We've trained models like Logistic Regression, Random Forest, Naive Bayes, and Decision Trees to classify Twitter accounts as malicious or benign. Our work contributes to enhancing social media security, protecting user trust, and combating fake accounts. Keywords-- Malicious Twitter bots, Machine learning, URL analysis, Social media security, Classification, Logistic Regression, Random Forest, Naive Bayes, Decision Trees, User trust, Fake accounts.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Detecting Malicious Twitter Bot using Machine Learning and URL analysis

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The purpose of this paper is to develop and evaluate a breast cancer prediction system using Convolutional Neural Networks (CNN) and Support Vector Machines (SVM). Within the wake of a breast cancer diagnosis, the project presents a comprehensive web-based platform created to meet the interrelated demands of administrators, healthcare providers, and patients. Patients are empowered to contribute to their healthcare journey by seamlessly uploading mammographic images. Through the integration of AI-driven algorithms, including CNN and SVM, the system predicts whether an individual has breast cancer or not can be determined by using the features that were taken from the images. For patients, a dedicated dashboard provides insights into diagnostic results, alongside some precautions. This paper examines the intricate integration of CNN and SVM algorithms, alongside patient- centric features and administrative controls, highlighting the potential impact on enhancing breast cancer diagnosis accessibility, efficiency, and overall user experience.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Support Vector Machine, Convolutional Neural Network, Machine Learning, Deep Learning, Cancer Prediction, Predictive Model.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In many domains, data duplication is a major problem that results in inefficient processing, storage, and analysis. In this work, we investigate the use of machine learning methods for tasks related to data duplication and deduplication, with a particular emphasis on textual and image data. This project provides preparation pipeline for textual data that includes tokenizing text, addressing missing values, and standardizing formats. Then, feature extraction techniques like word embeddings and TF-IDF are used to mathematically represent text. These attributes can be used to train machine learning models, such as Support Vector Machines (SVM) or clustering methods like K- means, which are used to efficiently identify and eliminate duplicate entries. These models are capable of identifying duplicate photos because they learn hierarchical representations of images. To successfully recognize duplicates, this entails creating hybrid models that integrate textual and visual information.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Data security, Deduplication, Authorization, Authentication, Access control, Cloud Security, Cipher Technology, Data synchronization

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In the era of information development, information transfer is far more practical. However, there is always a chance that the transmission method may be compromised, stolen, or assaulted, raising questions about the accuracy of the data source. Several academics suggested using passwords to secure sensitive data because of this. Proposed 3D-Playfair Cipher with Message Integrity Using MD5. The encryption used in this study is 3D-Playfair. Nevertheless, basic 3D-playfair encryption is unable to ensure the integrity of data while it is being sent, thus the author suggests In addition to using MD5 to guarantee data integrity, this study employs XOR calculation techniques to further confirm the data's legitimacy because there are concerns over the reliability of the data source.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Data security, Deduplication, Authorization, Authentication, Access control, Cloud Security, Cipher Technology, Data synchronization.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The diagnostic tropical cyclone intensity estimation system presented in this paper represents a novel approach to objective and automated intensity assessment using satellite imagery and deep learning techniques. Leveraging data from the National Hurricane Centre (NHC) and the National Oceanic and Atmospheric Administration (NOAA), the system integrates real- time storm outlooks with corresponding Geostationary Operational Environmental Satellite (GOES) imagery to provide accurate and timely predictions of cyclone intensity levels. The implementation involves web scraping for data retrieval, preprocessing of satellite imagery, and deployment of a convolutional neural network (CNN) model for intensity estimation. Results demonstrate the system's effectiveness in providing objective intensity assessments, with promising performance metrics such as Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE). Furthermore, the system's scalability and potential impact on disaster preparedness and response efforts are discussed, highlighting its role as a valuable decision support tool for meteorological agencies and emergency responders. Overall, this implementation paper contributes to advancing the field of cyclone intensity estimation and underscores the importance of integrating technology and data-driven approaches in disaster management.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Deep learning, wind speed estimation, Convolutional neural network, Estimation, Cyclone, Disaster management

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Tropical cyclones (TCs) stand as dynamic and complex atmosphere-sea interaction phenomena, their behavior contingent upon a delicate interplay of oceanic and atmospheric conditions. Their emergence, dissipation, or intensification presents a challenge for accurate prediction, and as such, the development of precise diagnostic models holds immense potential for saving lives and safeguarding property. Existing techniques for diagnosing tropical cyclone wind speeds have displayed varying degrees of success, often limited by their reliance on specific points in time and satellite data. This paper introduces a groundbreaking paradigm shift by presenting a deeplearning-based objective diagnostic estimation of tropical cyclone intensity. Leveraging the power of deep learning, the model promises to transcend the limitations of traditional methods, offering a more nuanced and accurate representation of cyclonic behavior. A key innovation lies in the integration of an infrared satellite imagery-based diagnostic system. This method represents a new visualization gateway in addition to improving intensity estimation accuracy. This portal is not merely a scientific tool; it stands as one of the first systems to seamlessly translate deep learning results into a user-friendly interface, presenting not only raw data but also contextual information to end users. This move towards user accessibility marks a significant stride in bridging the gap between advanced scientific methodologies and practical, realworld applications

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Deep learning, wind speed estimation, Convolutional neural network, Estimation, Cyclone, Disaster management

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The behaviors of cryptocurrency markets are dynamic and complicated, with high volatility and a wide range of influencing factors. In order to anticipate bitcoin prices, this study investigates the use of deep learning techniques, particularly recurrent neural networks (RNNs) and long short- term memory networks (LSTMs). For well-known cryptocurrencies, the study makes use of past price data, trade volumes, and extra technical indicators. The methodology includes feature engineering, training the model, evaluation, and preprocessing of the data. Measures like Mean Absolute Error (MAE) and Mean Squared Error (MSE) are used to evaluate the performance of the deep learning model. The results shed light on how well deep learning captures patterns and temporal relationships in cryptocurrency price data. The conversation explores how the findings could affect real-world trading tactics, points out its shortcomings, and suggests directions for further study. This study adds to the expanding corpus of research on the prediction of cryptocurrency prices by providing a sophisticated knowledge of the potential benefits and difficulties of using deep learning in this field.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Cryptocurrency, Price Prediction, Deep Learning, LSTM, Neural Networks, Financial Forecasting

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In this day and age the Mastercard extortion is the greatest issue and presently there is need to battle against the Visa misrepresentation. "Visa extortion is the most common way of cleaning messy cash, accordingly making the wellspring of assets as of now not recognizable." On consistent schedule, the monetary exchanges are made on gigantic sum in worldwide market and subsequently identifying charge card misrepresentation movement is testing task. As prior (Against Mastercard extortion Suite) is acquainted with distinguish the dubious exercises yet it is relevant just on individual exchange not for other financial balance exchange. To Conquers issues of we propose AI technique utilizing 'Underlying Closeness', to recognize normal credits and conduct with other financial balance exchange. Location of charge card misrepresentation exchange from huge volume dataset is troublesome, so we propose case decrease strategies to lessens the information dataset and afterward find sets of exchange with other financial balance with normal ascribes and conduct.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

credit card fraud, fraudulent activities, SVM (SUPPORT VECTOR MACHINE), Harr cascade Algorithm, Face Recognition.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The challenge of content-based image retrieval (CBIR) lies in its reliance on low-level visual features from user query images, making query formulation difficult and often yielding unsatisfactory retrieval results [1] . Previously, image annotation emerged as a promising solution for CBIR, employing automatic assignment of keywords to images for improved retrieval based on user queries. picture annotation essentially mirrors picture class, where low-level features are mapped to excessive-level principles (elegance labels) through supervised mastering algorithms. However, achieving effective feature representations and similarity measures remains critical for CBIR performance. The semantic gap, characterized by the disparity between machine- captured low-level image pixels and human-perceived high-level semantics, poses a significant challenge in this context. Recent advancements in deep learning, particularly Convolutional Neural Networks (CNNs), have demonstrated remarkable success in various computer vision tasks, motivating my pursuit to address the CBIR problem using a dataset of annotated images [6].

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Deep learning, Convolutional Neural Networks (CNNs), Similarity measures, Semantic gap, Computer vision

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The Contest Lister project aims to revolutionize the way coding enthusiasts engage with coding contests, hackathons, and hiring challenges across various online platforms. With the exponential growth of coding competitions, enthusiasts face the daunting task of keeping track of events, often leading to missed opportunities and disorganized participation[8]. In response to this challenge, the Contest Lister project presents a comprehensive solution that aggregates contest information from prominent platforms like CodeChef, HackerRank, and LeetCode into a centralized platform. By leveraging modern web technologies and APIs, the system provides users with real-time updates on upcoming contests, intuitive filtering options, and a user-friendly interface for seamless navigation. This paper outlines the architecture, methodology, results, and future prospects of the Contest Lister project, offering insights into its potential to streamline contest discovery and enhance user engagement in the coding community.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Coding contests, Competitive programming, Contest Lister, System architecture, API integration, User interface

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The Contest Lister project aims to revolutionize the way coding enthusiasts engage with coding contests, hackathons, and hiring challenges across various online platforms. With the exponential growth of coding competitions, enthusiasts face the daunting task of keeping track of events, often leading to missed opportunities and disorganized participation[8]. In response to this challenge, the Contest Lister project presents a comprehensive solution that aggregates contest information from prominent platforms like CodeChef, HackerRank, and LeetCode into a centralized platform. By leveraging modern web technologies and APIs, the system provides users with real-time updates on upcoming contests, intuitive filtering options, and a user-friendly interface for seamless navigation. This paper outlines the architecture, methodology, results, and future prospects of the Contest Lister project, offering insights into its potential to streamline contest discovery and enhance user engagement in the coding community.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Coding contests, Competitive programming, Contest Lister, System architecture, API integration, User interface

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The aim of this study is to provide a platform where SPPU university students can easily convert their cumulative grade point average (CGPA) to percentage. Usually, students are more comfortable in the percentage system as CGPA is a new concept to most of them. Percentage is the basic requirement criteria to be addressed/written in many official forms like job applications, scholarships forms, etc. As university students get their result in CGPA format, they need to convert them first in percentage, but the procedure of converting CGPA to percentage is not known to many students. There are many websites and formulae available on the internet, but the accuracy of each is different. In addition, the manual way of calculation can increase error rate, and may be hectic. To resolve this issue, a system is proposed that converts CGPA to percentage in an accurate format. A student needs to enter his/her CGPA, and in just one click, user gets a converted percentage. The system is crafted utilizing HTML, CSS, and JavaScript. The system can be executed in any operating system and is user- friendly.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

CGPA, Percentage, Grade, Class, User interface.

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

The process of image colorization has gained popularity, especially for converting old black and white photos into colorful ones, which provides a more immersive view of historical material. This process holds significant relevance for historical preservation and storytelling, providing a means to visualize the past with enhanced accuracy and realism. There are two primary approaches to image colorization: manual and automatic. Manual colorization relies on skilled individuals using software like Adobe Photoshop to meticulously add colors to grayscale images. This method demands expertise to ensure the chosen colors align with the historical context and period depicted in the image. In automatic colorization represents a newer advancement driven by deep learning technologies, such as convolutional neural networks. This technique determines the appropriate colors for a given image by analyzing its grayscale values. It frequently achieves these results with little assistance from a human. Moreover, preserving the original integrity of the image poses another challenge. While colorization can enhance visual appeal, it must be done sensitively to avoid altering the original meaning behind the image. Striking a balance between enhancing visual aesthetics and preserving historical authenticity is crucial to ensure that colorization efforts remain respectful to the original image's significance.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Convolutional Neural Network(CNN), Mean Square Error(MSE) ,peak signal noise ratio(PSNR),Structural Similarity(SSIM),Gray Scale , RGB Scale

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

If not caught and diagnosed early, brain tumors can turn into cancer. Today, the detection and classification of brain tumors is done by doing the biopsy, which can be a timeconsuming process. With less time and effort radiologists can now construct tumors thanks to advancement in technology and machine learning algorithms . First, are going to propose a model that will determine whether there are tumors in the brain by segmenting MRI images and, if they are detected, we use an architecture based on SVM to classify tumors in MRI images as tumors and no tumors function well. The foundation allows staff to decide on the repair process. The development of the model will be divided into a training and testing phases and will be tested using more data and more methods. The proposed SVM/KMEANS architecture achieves high accuracy, reliability and execution speed and will become a powerful diagnostic decision for radiologists.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Magnetic Resonance Imaging(MRI), Support Vector Machine (SVM), K- Nearest Neighbor(KNN).

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

Tumors present in the brain can develop into cancer if they are not discovered and treated quickly. These days, brain tumors are identified and categorized by the timeconsuming biopsy procedure. With less time and effort, radiologists can now construct tumors thanks to technological advancements and machine learning algorithms. Firstly, we provide a brain tumor detection model based on MRI image segmentation. If brain tumors are found, a deep learning- based SVM/KNN architecture is used to identify the tumors as well as their functional characteristics. The foundation gives personnel discretion over the repair procedure. The model will go through phases of training and testing, with additional methods and data being used for testing. With its excellent accuracy dependability, and speed of execution, the suggested SVM/KNN architecture will be a valuable diagnostic tool for radiologists.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

MRI ( Magnetic Resonance Imaging) , SVM( Support Vector Machine) , KNN( K Nearest Neighbour)

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

This study presents an innovative approach for accurately recognizing diverse Yoga poses using deep learning algorithms. Our proposed system introduces a method for Yoga pose assessment leveraging pose detection to facilitate self-learning of Yoga. Utilizing multi-parts detection solely with a PC camera, the system effectively identifies Yoga poses. Furthermore, we introduce an enhanced algorithm for scoring applicable to all poses. Evaluation of our application encompasses various Yoga poses across different scenarios, demonstrating its robustness. We propose a hybrid deep learning model incorporating linear regression for real-time Yoga recognition in videos. This model utilizes linear regression to extract features from key-points in each frame, obtained through Open-Pos.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Yoga Pose Recognition, Deep Learning Algorithms, Linear Regression, Real-time Video Analysis, Pose Detection System

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In this study, there is exploration of how people's BMI changed pre and post the onset of covid pandemic, considering factors like food habits (the nutritious value of what they eat) and physical activities. They not just eyeballing the data--system using super-smart computer techniques called Reinforcement Learning, specifically Deep Q Network and Random Forest Regression and Gradient Boost Regression. Before COVID-19, know people had certain eating food habits and lifestyle habits. Now, with the pandemic, those might have changed. Using Deep Q Network, our computer system learns from this data and figures out how these changes are linked to BMI. It's like teaching a computer to understand the consequences of different habits on weight. Gradient Boost Regression is another technique being used. It helps the computer learn not just from the data have but also by exploring possibilities like, what if someone changed their eating habits or exercise routines? This way, system not just looking at what happened but also predicting what could happen. By combining these techniques, study aim to unravel how food choices and physical activities during and after Covid-19 might have influenced BMI. It's like having a smart assistant to help us understand the connection between lifestyle changes and weight, shedding light on how they can stay healthy in this challenging time.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Body Mass Index (BMI), Random Forest Regression, Gradient Boosting Regression, Deep Q-Network (DQN).

  License

Creative Commons Attribution 4.0 and The Open Definition

 Abstract

In this study, there is exploration of how people's BMI changed pre and post the onset of covid pandemic, considering factors like food habits (the nutritious value of what they eat) and physical activities. They not just eyeballing the data--system using super-smart computer techniques called Reinforcement Learning, specifically Deep Q Network and Random Forest Regression and Gradient Boost Regression. Before COVID-19, know people had certain eating habits and physical activities. Now, with the pandemic, those might have changed. Using Deep Q Network, our computer system learns from this data and figures out how these changes are linked to BMI. It's like teaching a computer to understand the consequences of different habits on weight. Gradient Boost Regression is another technique being used. It helps the computer learn not just from the data have but also by exploring possibilities like, what if someone changed their eating habits or exercise routines? This way, system not just looking at what happened but also predicting what could happen. By combining these techniques, study aim to unravel how food choices and physical activities during and after Covid-19 might have influenced BMI. It's like having a smart assistant to help us understand the connection between lifestyle changes and weight, shedding light on how they can stay healthy in this challenging time.

Licence: creative commons attribution 4.0

  License

Creative Commons Attribution 4.0 and The Open Definition

 Keywords

Body Mass Index (BMI), Random Forest Regression, Gradient Boosting Regression, Deep Q-Network (DQN).

  License

Creative Commons Attribution 4.0 and The Open Definition