With the rapid development of the global economy and technology, Bitcoin trading has become an important channel for investors seeking wealth. However, its extreme price volatility presents significant risks and uncertainties, making it crucial to understand the driving factors behind these fluctuations. This research project analyzes Bitcoin price movements during the second half of 2024, a period characterized by dramatic price increases, particularly at year-end. Using a combination of visualization methods, ARIMA forecasting models, and isolation forest anomaly detection algorithms, the study examined the relationship between Bitcoin prices and major socioeconomic events. The findings revealed that the U.S. presidential election was the primary factor influencing Bitcoin’s significant price changes during this period. While technical analysis through ARIMA modeling provided valuable insights into price patterns, the strong correlation with political events demonstrated that major socioeconomic factors can substantially impact cryptocurrency markets. These results offer investors a more comprehensive framework for cryptocurrency investment decision-making, highlighting the importance of considering both technical indicators and external event impacts when navigating the volatile Bitcoin market.

With the rapid development of artificial intelligence(AI) and breakthroughs in computer deep learning and machine learning technologies, AI is gradually replacing traditional disease diagnosis. At the same time, medical institutions have also accumulated a large amount of case information, medical data, etc., providing powerful and rich databases for AI disease prediction and diagnosis, thus achieving more accurate prediction and diagnosis of AI. AI has been widely applied in the medical field and continues to make new progress in areas such as medical image analysis, biological signal analysis, and case analysis. This article reviews the research progress and related technologies of AI in medical assisted diagnosis and disease prediction in recent years, aiming to promote research in the field of artificial intelligence medicine. AI has made significant achievements in the fields of medical assisted diagnosis and disease prediction. Medical image and biological signal analysis have greatly improved the accuracy of disease diagnosis, but still face many challenges.

With the rapid development of computer technology, cloud computing has been widely used in all walks of life, but the consequent network security problems have become increasingly prominent. This paper introduces data encryption technology to study the security challenges in cloud computing. Firstly, the basic characteristics of cloud computing are analyzed, and the security vulnerabilities of cloud computing are discussed. This study examines the function of data encryption technology in cloud computing in light of these issues, and deeply analyzes the applications and solutions of technologies for symmetric and asymmetric encryption, technologies for link and end-to-end data encryption both and node data encryption technology, in order to provide a reference for improving data security in cloud computing environment. The proposed data encryption scheme not only ensures data security but also effectively reduces the impact of encryption on the performance of cloud computing, demonstrating high practicality and feasibility. The research in this paper offers fresh concepts and techniques for the field of cloud computing security and is of great significance for promoting the wide application of cloud computing technology.

This review paper focuses on expanding application scenarios of consensus in multi-agent systems (MASs), particularly in smart manufacturing, intelligent transportation, and public safety. It provides a comprehensive analysis of recent methods for solving consensus problem. The paper focuses on various control, including distributed control, optimization-based consensus, event-triggered consensus control, game-theoretic consensus, and novel approaches based on fundamental theoretical research. The paper summarizes the main challenges in achieving consensus and highlights future research directions aimed at overcoming these challenges. By offering a comprehensive overview and analysis, this review paper aims to assist researchers and practitioners in selecting the optimal method for practical applications and fostering further advancements in consensus.

The 21st century is the century of information explosion. People around the globe were gradually shifting from bulky desktop computers to using portable smartphones to demand faster and more accurate information transmission. The purpose of this paper is to discuss the definition and development of networks, from 4G to 6G, exploring the applications supported by networks in different eras and how they gradually change netizens lives. It reveals that while 4G only supports smooth video, 6G is expected to be gradually implemented by 2030, which will lead to the normalization of big data analysis and XR usage, thereby achieving a fully connected society among ocean, land and space. This article further analyzed the implementation of 4G, 5G, and 5.5G in global cities and rural areas, revealing that the theoretical values and actual performance of 4G to 5G, 5.5G are indeed disappointing. It is worth considering that while there is still significant room for improvement from 4G and 5G to 5.5G, the performance of 6G in people's lives should also be just as unsatisfactory under such urgent time and financial budgets.

With the rapid development of the social economy, significant improvement in the living standards of residents, changes in lifestyle, and the increasingly severe issue of population aging, cardiovascular diseases have become one of the most widespread diseases, posing a serious threat to public health. In order to explore the inherent patterns of the occurrence and development of these diseases, it is necessary to use quantitative methods to describe the correlation between risk factors and disease incidence, as well as to predict the epidemic trends of the disease. This will provide scientific theoretical support for medical professionals and public health prevention agencies, enabling the implementation of effective preventive and control measures. In the field of dynamic prediction of non-stationary disease incidence, various methods have been proposed, each with its own advantages and disadvantages. Artificial neural network technology, particularly the classic BP neural network, has shown significant advantages in handling nonlinear pattern recognition and prediction tasks. Therefore, this study constructs a disease incidence prediction model based on the BP neural network and verifies the model’s effectiveness through the analysis of practical cases.

This research addresses the critical challenge of energy-efficient anomaly detection in resource-constrained edge computing environments for IoT networks. With the proliferation of IoT devices generating exponential data volumes and information technology energy consumption projected to reach 20% of global electricity production by 2030, sustainable computing approaches at the network edge are imperative. We propose a novel optimization framework that dynamically balances computation offloading decisions with local processing capabilities to minimize energy consumption while maintaining detection accuracy and meeting real-time requirements. The framework incorporates: (1) a calibrated energy consumption model for heterogeneous edge environments, (2) an adaptive resource allocation strategy responding to network conditions, (3) lightweight machine learning architecture optimized for minimal energy footprint, and (4) intelligent computation offloading based on device energy states. Experimental evaluation on a testbed of 16 heterogeneous edge devices processing real-world IoT traffic demonstrates energy consumption reduction of 23.8% compared to traditional approaches, while maintaining detection accuracy above 92.5% across diverse anomaly types. The system extends battery life by up to 165% in energy-constrained scenarios through dynamic adjustment of detection parameters. Comparative analysis confirms superior performance against state-of-the-art methods in both energy efficiency and detection capability, particularly in environments with variable energy availability.

Music is an integral part of life, and in recent years, machine learning has been increasingly applied to music generation. While many existing review papers focus on the models used and the overall development of the field, few have examined the popularity of different musical genres within AI-generated music. This paper aims to fill that gap by analyzing the popularity of AI-generated music from two angles: overall trends and genre-specific trends. It also offers possible explanations for the patterns observed. To assess popularity, the number of related publications indexed on Google Scholar each year is collected and analyzed. The paper is divided into two main sections. The first examines the general rise in interest in AI-generated music, particularly the sharp increase after 2016. This surge may be attributed to developments such as more efficient models, improved and affordable hardware, and broader access to training data. The second section focuses on the three most popular genres in AI music generation, comparing their relative popularity and exploring the reasons behind these trends. Factors such as data availability, genre complexity, and real-world popularity are discussed as possible explanations.

Due to the high volatility of financial markets and the prevalence of financial fraud, real-time stock market forecasting for listed companies remains a challenging task. To address these challenges, this study proposes TCNAttention-RAG, a hybrid deep learning framework integrating Temporal Convolutional Network (TCN), Multi-Layer Perceptron (MLP), Attention Mechanism, and Retrieval-Augmented Generation (RAG) for enhanced stock price forecasting. The model leverages TCN for temporal feature extraction, MLP for nonlinear representation, and Attention for feature weighting, while RAG dynamically retrieves key financial insights from corporate reports to improve predictive accuracy. Using NASDAQ-listed stock price data (2014–2020), combined with corporate financial reports, market transaction data, and macroeconomic indicators, a multi-dimensional dataset is constructed. Experimental results demonstrate that TCNAttention-RAG outperforms traditional models in accuracy and recall, effectively capturing stock price fluctuations. Despite its limitations in handling extreme market events, the model exhibits high reliability and predictive robustness. This study introduces a multi-modal data-driven approach to financial forecasting, offering insights into intelligent financial analysis and enhancing decision-making in volatile markets.

Brain tumors, as central nervous system diseases that endanger human health , require early and accurate detection to significantly improve patient survival rates. Magnetic Resonance Imaging (MRI), due to its superior soft-tissue contrast and non-invasiveness, has become a crucial tool in brain tumor diagnosis. However, traditional imaging diagnosis, which heavily relies on manual interpretation, suffers from limitations such as strong subjectivity and low efficiency. To address these issues, this paper proposes an automatic brain tumor detection method based on Convolutional Neural Networks (CNN). By leveraging deep learning techniques, the method extracts multi-level features from MRI images to achieve high-precision classification of glioma, meningioma, pituitary tumor, and non-tumor categories. A lightweight CNN model was developed, incorporating data augmentation and normalization preprocessing strategies. Experiments were conducted on a dataset of 7,023 MRI images. The results show that the model achieved classification accuracies of 96% on the training set and 95% on the validation set, demonstrating strong robustness and generalization capability. Confusion matrix analysis indicates that the model maintains high recognition accuracy across all categories, with particularly outstanding performance in identifying non-tumor and pituitary tumor cases. This study provides an effective technical pathway for intelligent assisted diagnosis of brain tumors and holds promising clinical application potential.