As a cryptocurrency, Bitcoin serves as a decentralised ledger for recording transactions. The owner of a Bitcoin keeps their identity secret and hides it behind a special address known as a pseudonym. Because Bitcoin offers anonymity, it has evolved into the favoured option for cybercriminals involved in illegal activities. In this research, supervised machine learning has been used to propose a framework for identifying anonymous user activities on the Blockchain. A labelled dataset containing transactions has been created as a training dataset to carry out the classification of user activities. The fundamental objective is to classify Blockchain transactions to deanonymise them and separate unethical from ethical ones. Synthetic minority oversampling technique (SMOTE) and weight of user activities were used to address the issue of class imbalance. On the samples from the class imbalanced and class balanced datasets, k-nearest neighbours (KNN) exhibited outstanding cross-validation accuracy with default parameters and hyperparameters.

Periodic Leg Movement during Sleep (PLMS) and Bruxism are linked with changes in EEG signal characteristics. This work applies machine learning and data mining approaches to examine these changes. Patients with PLMS and bruxism had nighttime EEG recordings to examine changes in brain activity. The findings revealed constant variations in brain hemodynamics even in the absence of clearly observable arousals in the EEG. Wavelet decomposition was used to improve classification precision. Using the N3 sleep stage, accuracy varied from 92 % to 96 %, with an AUC of 0.85–0.89, in diagnosing binary sleep disorders. Still, adding wavelet-based elements greatly enhanced performance, obtaining an AUC of 0.99 with classification accuracy ranging from 94 % to 98 %. This emphasizes how strongly discriminative power wavelet-extracted EEG characteristics possess. Using K-Nearest Neighbors (KNN), Artificial Neural Networks (ANN), and Support Vector Machines (SVM) with Radial Basis Function (RBF), Bruxism categorization was accomplished. These models attained respectively 82 %, 90 %, and 93 % percent classification accuracy. This work is the first to show a direct connection among differences in brain activity based on PLMS, Bruxism, and EEG-based technologies. The results show how well machine learning methods and EEG feature extraction might diagnose sleep problems. Although the therapeutic relevance of these findings is yet unknown, the results imply that enhanced EEG-based classification techniques could produce more reliable and automated diagnostic instruments for Bruxism and PLMS.

In the dynamic field of cybersecurity, safeguarding network integrity is paramount. IDS plays a pivotal role in this endeavour but faces challenges such as decreased detection rates, accuracy issues, and false alarms. Machine learning techniques can help by telling the difference between normal and abnormal network behavior. Deep learning is especially good at this because it can pull out the most useful traits from raw data. This essay looks at how Machine Learning techniques can be used to make Intrusion Detection Systems (IDS) work better. Some of the methods that were looked into are Decision Tree (DT), Support Vector Machine (SVM), Logistic Regression (LR), Random Forest, XG Boost, Gradient Boosting, and Light GBM. In their experimental work, authors have shown the algorithm accuracy as 99

An effective strategy for identifying and categorizing toxic comments is urgently needed due to the overwhelming number of such comments on various internet platforms. While previous research has addressed various forms of toxic behaviour, existing approaches often fall short. Our research fills this need by suggesting a novel method that combines ML with natural language processing (NLP). Building on previous work, we focus on recurrent neural networks (RNNs), specifically LSTM networks, renowned for their adeptness in capturing sequential patterns and contextual information within text. Also, we incorporate dropout layers to enhance model generalization and combat overfitting. Evaluating our models on a benchmark dataset reveals promising outcomes, with the LSTM model achieving an accuracy score of 0.9943 and a loss of 0.0493, while the LSTM model with dropout layers closely follows with an accuracy score of 0.9939 and a loss of 0.0493. These findings underscore how effectively our methodology detects toxic comments, thereby advancing automated moderation systems. Looking ahead, we aim to refine our approach and explore avenues for real-world implementation, aiming to foster healthier online discourse and enhance user safety.

In contemporary agriculture, maximizing crop growth and minimizing risks are critical concerns, especially within sugarcane cultivation. This research work introduces a monitoring system tailored explicitly for sugarcane crops to tackle these challenges. Previous research underscores the potential of precision agriculture systems in enhancing crop yields and managing resources effectively. Integrating advanced data analysis techniques with real-time sensor data, the methodology monitors essential environmental parameters. Passive infrared sensors are integrated to identify and mitigate potential threats to crop health. The results exhibit promising accuracy, with naive Bayes achieving 96

Rainfall is a key aspect in agriculture, which is important to the Indian economy. Rainfall forecasting, however, has grown more challenging in recent years. Accurate forecasts can assist farmers in making better plans for their crops and in taking the appropriate measures. The effects of global warming are also seen in the changing climatic conditions, which have an impact on both humans and nature. Floods, droughts, and unpredictable and excessive rains are all being brought on by the warming of the atmosphere and rising ocean levels. For the different types of industries, comprising agriculture, research, and energy production, the ability to predict rainfall is crucial since it aids in understanding climate change and its associated variables, including temperature, humidity, precipitation, and wind speed. In order to create a more precise rainfall prediction system, the objective of this research piece is to employ multiple machine learning classification techniques such as Random Forest Classifier, Gaussian NB, K-Neighbors Classifier, and XGB Classifier, to a dataset obtained from the Kaggle repository that contains multiple features. These algorithms were ranked in order of efficiency, with Random Forest Classifier and XGB Classifier giving highest accuracy.

Today, data analytics has become quite useful for everyone. Using the right tools, decision-makers can extract different insights of data which could help to optimize various parameters. It can help to recognize several patterns from the given loads of information. These kinds of insights can be quite useful to increase the overall performance of any company. According to this research and looking at the future scope of data science and analytics field, our team has used Qliksense software for creating an analytical tool known as “E C AniTool” which is an advanced kind of dashboard consisting of various features and parameters for doing advance evaluation and analysis of any company data. This analytical tool can help manage the big data on a daily basis for any company using various visualization techniques which can be quite useful for any company. Qliksense is a software for visualizing data and discovering many products/tools that allow us to create user-friendly and communal visualizations that lead us to meaningful processes and decisions. The main reason to use Qliksense is that it can be accessed anywhere, anytime, and through any device. It is a modern data analytical software which is being used by many multinational corporations.

In response to the escalating concern over road accidents due to driver drowsiness, this research delves into enhancing drowsiness detection systems using advanced deep learning techniques and transfer learning. Prior investigations have explored conventional methods, yet they often lack the precision demanded by real-world scenarios. Leveraging convolutional neural networks (CNNs) alongside model fine-tuning and transfer learning, our methodology aims to optimize drowsiness detection accuracy. Through rigorous experimentation, we achieved compelling results, showcasing significant enhancements in accuracy across various CNN architectures. For instance, with Max Pooling and 2 Dense layers, one of our CNN models achieved an accuracy score of 1 at epoch 2, with a remarkably low loss of 6.622934e-12, all while taking only 2.6648 seconds for training. These outcomes underscore its potential to revolutionize driver safety systems, offering promising avenues for future research and implementation in real-world contexts.

This is a comprehensive exploration of Human Activity Recognition (HAR) and its integration with deep learning techniques, specifically Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks. CNNs are widely used in HAR by extracting spatial features; however, CNNs are not competent of extracting global features. LSTM network evaluates sequence data by looping across time steps, and identify long-term dependencies between time steps, which plays a crucial part in HAR. To address this concern this study primary objective is to develop a CNN-LSTM deep neural network-based HAR model. The paper begins with an introduction outlining the objectives and constraints of the research. A detailed literature review follows, delving into the fundamentals of HAR, its relationship with machine learning, and the introduction of CNN and LSTM in this context. The methodologies section covers data gathering, preprocessing, model architecture, hyperparameter tuning, and the learning and evaluation processes. The challenges within the field of HAR, such as intraclass differences, interclass similarity, and recognition in real-world scenarios, are discussed in the subsequent section. The results of the model execution, including the performance of the CNN + LSTM model and a comparison with baseline models, are presented and discussed. This paper concludes by highlighting the key findings and their real-life use case scenarios, and it discusses the potential for future research in the field of HAR. In summary, this paper aims to contribute to the understanding and advancement of HAR models, particularly in complex, diverse, and real-world scenarios. The proposed methodology enhances the recognition accuracy with value of 98.67

Bitcoin is a digital currency that provides a way to transact without any trusted intermediary; however, privacy is an issue. Numerous deanonymization endeavors have been proposed, in spite of the fact that Bitcoin addresses aren’t linked with a specific identity. In this work, blockchain transactions are deanonymized using ensemble learning. An excess of four million labeled dataset samples comprising user activities such as pools, services, gambling, and exchanges have been gathered from various repositories and prepared for training and validation to perform the classification. The main aim is to deanonymize blockchain transactions via classification and separate legitimate ones from illegitimate ones. On the class imbalanced dataset, remarkable cross-validation accuracy was attained using the EXtreme Gradient Boosting with default parameters and hyperparameters. Using EXtreme Gradient Boosting, Random Forest, and Bagging on the class-balanced dataset produced the best cross-validation accuracy when using the default parameters and hyperparameters. The empirical findings indicate that the effectiveness of the proposed deanonymization using the proposed ensemble learning model has achieved up to 98.45

Social media platforms, especially Twitter, have become a popular channel for customers to express their grievances related to products or services. Companies need to have an efficient grievance redressal system in place to address these complaints in real-time to ensure customer satisfaction and loyalty. This research paper presents the development of a cloud-based integrated real-time Twitter grievance redressal system using Amazon Web Services (AWS) and machine learning approach. The proposed system uses AWS cloud infrastructure to lever the huge volume of dataset created by tweets and machine learning algorithms to classify and prioritize them based on their severity. The system includes a web application that allows the grievance redressal team to view, categorize, and respond to the tweets efficiently. The efficiency of the planned system is evaluated via a case study. The results show that the system can effectively handle a huge volume of tweets and improve the grievance redressal process. The system’s response time is significantly reduced, and the team can prioritize the tweets based on their severity and importance, leading to better customer satisfaction. Data security is a critical aspect of the proposed real-time application as it will be handling sensitive data of the users. Therefore, security measures such as encryption, MFA, and disaster recovery must be properly implemented and configured, in order to ensure the security of data of the suggested grievance redressal system. The suggested system has achieved an accuracy of 89.5

Periodic leg movements (PLM) and bruxism are prevalent sleep disorders that significantly impact sleep quality and overall health. Accurate and timely detection of these disorders is pivotal for effective treatment and improved patient well-being. In this research paper, we present a novel approach for the identification of Periodic Leg Movements (PLM) and bruxism. Our methodology leverages sophisticated complexity and entropy features extracted from electroencephalogram (EEG) signals recorded across various sleep stages. The implementation involves a diverse set of artificial intelligence (AI) classifiers, encompassing K-nearest neighbours (KNN), artificial neural networks, support vector machines (SVM) with radial basis function, SVM with polynomial kernel, and SVM with sigmoid kernel. These classifiers are employed to categorize EEG signals into three distinct classes: PLM, bruxism, or normal sleep patterns. To underscore the efficacy and superiority of our proposed approach, we conduct a thorough comparison with existing publications in the field. Our approach consistently outperforms previous methods and attains remarkable accuracies of 98.74% for PLM and 94.74% for bruxism, particularly when utilizing the SVM Sigmoid classifier. Furthermore, our study highlights the potential of EEG signals in elucidating alterations in brain activity associated with PLM and bruxism, underscoring the reliability of complexity and entropy features as robust biomarkers for these sleep disorders. Our research significantly advances the contemporary pinnacle in EEG-based detection of Periodic Leg Movements (PLM) and bruxism signifies a significant advancement, holding potential for impactful clinical applications in the realm of sleep disorder diagnosis and treatment.

Many frameworks with packaged pre-guided models have been created to give users fast access to transfer learning considering the rapidly expanding field of object detection techniques. For instance, three well-known computer vision systems with trained models are GluonCV, Detectron2, and the Object Detection API for TensorFlow, as well. The TensorFlow2 Object Detection API is an update to the TensorFlow Object Detection API. One of the cutting-edge object identification algorithms that can be trained using the TensorFlow2 Object Detection API is EfficientDet from Google Brain (Implemented here). Authors created and implemented an EfficientDet model with the help of TensorFlow Object Detection API using Amazon SageMaker. It is constructed on top of TensorFlow 2, which facilitates its creation, training, and deployment. Because it is designed on top of TensorFlow 2, creating, training, and deploying object detection models is simple. SageMaker is a completely managed tool that lets data scientists and developers quickly build, direct, and implement ML models. To make it simpler to create high-quality models, SageMaker takes the labour-intensive tasks out of each stage of the ML process. Transfer learning on numerous pre-guided models accessible in TensorFlow Hub is made possible by object detection with TensorFlow in SageMaker. The head of the TensorFlow model that handles object detection is replaced based on the amount of class labels present in the guiding data. Based on fresh guiding data, either the entire network—including pre-guided model—or just the top layer (object recognition head) can be fine-tuned. Authors trained using a smaller dataset in this transfer learning method. Authors have talked over each step-in detail, including data collection and labelling with Ground Truth, making, and converting the data to TFRecord format, training as well as launching a special object detection model with the TensorFlow Object Detection API, and ultimately deploying the model.

It has been demonstrated that periodic leg movements during sleep (PLMS) are connected to alterations in features of EEG signal. Data mining evaluates hemispheric/cortical activity-related hemodynamic changes. We used data mining and machine learning to examine whether there are changes in brain hemodynamics associated with PLMS. Nighttime EEG recordings were made while brain activity was monitored in PLMS patients. Scores from EEG feature data were examined to find relevant differences. PLMS were consistently accompanied by variations in brain activity that increased in magnitude when related to changes in EEG and persisted even in the absence of any arousal that could be seen visually in the EEG. This study is the first to show PLMS-related alterations in brain activity. Clinical relevance of these observations has yet to be established. We have used wavelet decomposition with or without it to complete the aforementioned classification tasks. For binary disease to identify tasks employing the sleep stage of N3, we have achieved classification accuracy ranging between 92% and 96% and AUC ranging between 0.85 and 0.89. But, as the use of the suggested wavelet-based features is performed, a superior classification accuracy is achieved, with an AUC of 0.99 and a range of 94–98%. This is because the best wavelet-based features have a high degree of discrimination.

Due to internet usage, people have experienced rapid changes. The fact is that the internet helps people to manage their social networks and links for their assistance whenever they require it. When users connect to the internet, they share their personal and professional data across the network with several risks with each other or any organization. As all know the internet is a crucial aspect of an individual’s day-to-day life, our data could be at risk at any time. For this purpose, the concept of intrusion detection system (IDS) may be proposed. IDS is important for protecting internet users against malicious or unauthorized attacks. An intrusion detection system monitors suspicious activity in network traffic, if it finds any issues or illegitimate activity, it immediately issues an alert message. In this paper, the main focus will be on different machine learning classification techniques; starting with machine learning algorithms KNN, logistic regression, decision tree, etc. These algorithms are used for finding the best accuracy or other performance metrics using the NSl-KDD datasets in the first step. Based on the first step’s result, the second process is the database with the most versed algorithm. Authors have considered NSL-KDD datasets for experiments and evaluating the model performance, and it ensures the efficient efficiency of our proposed model. The main focus was to improve the accuracy level of the model. Experiments on datasets implemented demonstrate different classifications through which authors find that decision trees provide the best result.

Object identification is a well-known research subject in the field of computer vision, with various applications like surveillance, autonomous driving, and robotics. The integration of machine learning with cloud computing has enabled organizations to automate many procedures and tasks, cut costs, and boost efficiency. With the help of a wide range of machine learning (ML) services offered by cloud computing platforms like Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP), organizations may take advantage of ML's potential without the need for specialized equipment or costly staff. A cloud-based ML service called Amazon Rekognition offered by Amazon Web Services is a powerful tool for object identification. Through this paper, the authors offer a study on the application of Amazon Rekognition for object detection and recognition. The idea is to detect objects in the provided images using machine learning and deep learning algorithms provided by Amazon Rekognition. The effectiveness of Amazon Rekognition in recognizing objects in images is precisely examined by the authors, who compare the discovered objects with state-of-the-art object detection algorithms and then provide the result with a corresponding confidence percentage. Experimental results show that Amazon Rekognition handles object detection tasks well, achieving a good balance between accuracy and speed. It is an effective tool for object detection with high average precision and recall values for many object categories. However, accuracy may vary depending on the complexity of the objects in the image, the lighting conditions, and other factors. Amazon Rekognition is a managed service that makes use of encryption, access control, compliance, monitoring, and logging. While the infrastructure and security are handled by AWS, it's crucial to incorporate security best practices within the application for maximum security. It is important for developers to carefully evaluate the performance of Rekognition for their specific use case and adjust the parameters and algorithms accordingly.

EEG data has proved to reflect the activities of the brain over all the sections with respect to human activities. It is useful in both cases that are in awakening states and in the sleep stage. Most of brain disease is due to the deterioration of brain cells. In this research work, EEG data is used to generate some frequency-based energy level features for NFLE (Nocturnal Frontal Lobe Epilepsy) patients and normal persons. After experimentation, it is found that the percentage of energy levels may be considered as a parameter to distinguish healthily and defected EEG data due to sleep disorder. During this study, it is also found that S0 is proved to be helpful in the diagnosis of NFLE sleep disorder because in this stage the percentage energy level is to be high for alpha waves for patients having a sleep disorder. The results show that the average percentage energy level in NFLE patient is 0.009872 while in normal cases it is 0.0010148.

Blockchain is one of the most disruptive and emerging technology in the present era, due to its unique characteristics immutability and decentralization. The transactions in a blockchain are publicly available, thereby accessible to all users. So, the privacy provisioning in blockchain for transactions of decentralized social media as well as privacy of the identities of members is needful. The chapter provides a comprehensive study of privacy provisioning on the blockchain. In the existing literature, there are a variety of scalability, security, and privacy concerns with blockchain, including key management for recovery, on-chain privacy protection, transaction linkability, and adherence to privacy laws. Following the self-sovereign identity (SSI) model, new privacy-preserving blockchain solutions are emerging to address these issues. These solutions allow users to anonymize themselves in all types of electronics transfer ledgers and regain control of their personal information. This chapter provides a comprehensive study of privacy provisioning schemes along with their challenges in this fascinating and disruptive technology.

Ethereum is a digital asset whose transactions are kept on a decentralized, globally accessible ledger. An Ethereum Blockchain owner's real identity is concealed behind a pseudonym termed an address. Because of this, Ethereum is frequently used in illegal activities like gambling and ransomware attacks because it is popularly believed to offer the highest level of anonymity. As a result, it is necessary to categorize the various malicious cybercriminal users' activities and addresses in the Ethereum Blockchain. The Blockchain's public data enables an in-depth analysis. Using supervised machine learning models including linear, non-linear, and ensemble learning models based on malicious and non-malicious activities, the classification of Ethereum Blockchain addresses is carried out in this paper. In this research work, cross-validation accuracy, recall, precision, and f1-score have been employed for the assessment. Findings indicate that linear and non-linear machine learning approaches are superior to ensemble learning for classifying Ethereum Blockchain addresses. The results also show that it is possible to discover the Ethereum Blockchain addresses of malicious users.

A public ledger is used by Bitcoin, a digital currency, to keep track of transactions. The owner of the Bitcoin keeps their identity secret and is identified only by their unique address. This indicates that because Bitcoin offers anonymity, it may be utilized for illicit purposes on a regular basis. This study presents a supervised machine learning approach for predicting anonymous user activities on the Bitcoin Blockchain. As a training dataset to facilitate the user activities classification, we created a labelled dataset with over 4 million samples from exchanges, gambling, pools, and services whose identities and types were disclosed. The primary goal is to classify transactions on the blockchain in order to deanonymize them and distinguish between legitimate and illegitimate ones. On the class imbalanced dataset, we obtained impressive cross-validation (CV) accuracy using the Gradient Boosting, Random Forest, and eXtreme Gradient Boosting with default parameters and hyperparameters. Using Random Forest helped achieve the best cross-validation accuracy on default parameters and hyperparameters obtained using grid search on the class-balanced dataset using the Synthetic Minority Oversampling Technique, while Bagging and eXtreme Gradient Boosting were used on hyperparameters obtained using randomized search. Empirical results show that the recommended model is up to 98% accurate.