Опубликована: Март 25, 2024
Здобувач факультету природничої
Academic Journal of Science and Technology, Год журнала: 2024, Номер 10(1), С. 299 - 304
Опубликована: Март 26, 2024
With the rapid development of internet technology, many industries have embarked on a digital transformation. However, while Internet has brought convenience to users, it also become breeding ground for criminals commit fraud. On one hand, large number users more or less left data, can use this information practice accurate fraud improve success rate fraud; other online financial transactions such as banking and e-commerce provide opportunities Therefore, all kinds methods emerge in an endless flow, through telephone, information, fishing means fraud, not only bring hundreds millions losses society every year, but security people's lives huge threat. Monitoring preventing is important part cybersecurity industry. For known network based domain name phishing site, account mobile phone that send fraudulent simple effective monitoring defence be carried out blacklist. difficult traditional effectively defend against undocumented machine learning main research direction detection discover sources characteristics content make real-time continuous judgments. This paper realises credit by generating adversarial so prevent risks.
Язык: Английский
Процитировано
32
Academic Journal of Science and Technology, Год журнала: 2024, Номер 10(1), С. 33 - 39
Опубликована: Март 26, 2024
In today's increasingly digital financial landscape, the frequency and complexity of fraudulent activities are on rise, posing significant risks losses for both institutions consumers. To effectively tackle this challenge, paper proposes a machine learning-based K-means clustering method to enhance accuracy efficiency fraud detection. By vast amounts transaction data, we can identify anomalous patterns behaviors in timely manner, thereby detecting potential fraud. Compared traditional rule-based detection methods, approaches better adapt ever-evolving techniques while improving flexibility precision Moreover, also aids optimizing resource allocation within by enabling focused monitoring prevention efforts high-risk areas, thus mitigating impact overall system. summary, holds promising prospects application field as it strives establish more secure reliable environment finance industry.
Язык: Английский
Процитировано
31
Academic Journal of Science and Technology, Год журнала: 2024, Номер 10(1), С. 62 - 68
Опубликована: Март 26, 2024
Natural Language Processing (NLP) is an interdisciplinary field of computer science, artificial intelligence, and linguistics that focuses on the ability computers to understand, process, generate, simulate human language in order achieve have natural conversations with humans. The underlying principles processing are at multiple levels, including linguistics, statistics. It involves study structure, semantics, grammar pragmatics, as well statistical analysis modeling large-scale corpora. In process concrete implementation, it necessary levels. Based this, this paper combined deep learning technology conduct sentiment patients' comments, so recommend drugs more suitable for patients, thus achieving accurate drug prescribing personalized recommendation.
Язык: Английский
Процитировано
19
Journal of Theory and Practice of Engineering Science, Год журнала: 2024, Номер 4(03), С. 1 - 8
Опубликована: Март 19, 2024
At a time when artificial intelligence is widely used in all walks of life, the way users interact with digital world also needs to incorporate intelligent elements reduce cost connectivity. This can be quantified through "experience metrics", which reveal problems encounter using interface (UI), and make targeted optimization. With AI, deep learning prediction user behavior achieved anticipate address potential barriers use UI design. will not only improve experience, but promote development design more user-friendly direction. Through accurate analysis experience indicators combined AI technology optimize design, gap between greatly reduced, making products suitable for achieving seamless interactive experience.
Язык: Английский
Процитировано
18
Journal of improved oil and gas recovery technology., Год журнала: 2024, Номер 7(3), С. 15 - 22
Опубликована: Май 15, 2024
This article reviews the key role of distributed cloud architecture in autonomous driving systems and its integration with intelligent computing networks. By spreading resources across multiple geographic locations, enables localized processing storage data, reducing latency improving real-time decision making vehicles. The points out that combination technology network provides a powerful solution to meet challenges technology. dynamically allocating deeply integrating cloud, network, chip technologies, gives enhanced data capabilities ensure stable reliable performance variety scenarios. Finally, paper highlights synergy marks an important milestone for transportation systems, heralding accelerated adoption solutions automotive industry, pace innovation transformation.
Язык: Английский
Процитировано
17
International Journal of Computer Science and Information Technology, Год журнала: 2024, Номер 2(1), С. 52 - 58
Опубликована: Март 13, 2024
The advent of Artificial Intelligence (AI) and Machine Learning (ML), particularly deep learning, has escalated the demand for computing resources. However, high hardware requirements pose challenges companies, compelling them to outsource ML tasks cloud. Nevertheless, concerns about cloud trustworthiness limit such applications. Encrypting data before uploading it is a straightforward solution ensure security. traditional encryption schemes render ciphertext unable participate in operations within domain, posing analysis. This paper delves into pivotal role homomorphic addressing critical issue privacy protection machine learning.
Язык: Английский
Процитировано
13
Academic Journal of Science and Technology, Год журнала: 2024, Номер 10(1), С. 290 - 298
Опубликована: Март 26, 2024
Machine learning is a branch of artificial intelligence (AI) technology that enables systems to learn and make predictions decisions without the need for explicit programming. algorithms patterns relationships from data are able gradually improve their accuracy.This paper mainly introduces application deep in financial academia industry, with particular focus on machine quantitative trading. The mentions complexity challenges markets difficulties faces processing time series data, such as overfitting, non-stationarity, heteroscedasticity autocorrelation. To overcome these challenges, explores model design trading risk prediction, including transformation construction. In related work part, it history development investment, well high-frequency trading, arbitrage strategy commodity advisory strategy. It then focuses quantification, use supervised unsupervised stock price forecasting generation. also compares traditional strategies discusses advantages solving high-dimensional non-linear problems. methodology section random forest principle, feature importance calculation experimental design. Through pre-processing, selection construction credit prediction dataset, evaluation process demonstrated. Finally, performance evaluated compared other models demonstrate its applicability quantification.
Язык: Английский
Процитировано
13
Academic Journal of Science and Technology, Год журнала: 2024, Номер 10(1), С. 50 - 55
Опубликована: Март 26, 2024
Human behavior recognition refers to the classification task of identifying specific actions human characters based on characteristics body and completed through a algorithm. It has wide range applications in intelligent surveillance, video retrieval so on. The main challenge this direction is accurately extract semantic information each describe its dynamic changes space time. Therefore, article introduces latest research progress field recognition. Through deep learning techniques, particularly convolutional neural networks recurrent networks, movements data can be effectively identified. However, models lack interpretability, which practical applications. researchers also introduce application traditional methods learning-based recognition, explore advantages processing multi-time scale introducing attention mechanisms. Finally, paper summarizes potential technology combined with multimodal behavioral analysis, provides prospects for smart fitness, health care other fields.
Язык: Английский
Процитировано
9
International Journal of Computer Science and Information Technology, Год журнала: 2024, Номер 2(1), С. 348 - 358
Опубликована: Март 24, 2024
A data warehouse is a subject-oriented, integrated, relatively stable collection of that reflects historical changes and used to support management decisions. Common tools for building are IBM Cognos SAP BO. However, both the above use centralized single-node mode build warehouses. This type has poor scalability, due rapid increase in scale Internet, traditional warehouses can no longer meet actual needs use. paper mainly introduces integration cloud machine learning as well importance application parallel methods. First, describes how combination warehousing promote business innovation output. It then discusses challenges managing models production environments, role addressing these challenges. Subsequently, computing Snowflake, implementation steps processes approach also introduced detail. Finally, results method analyzed, it considered good prospect development potential warehouse.
Язык: Английский
Процитировано
8
Journal of Theory and Practice of Engineering Science, Год журнала: 2024, Номер 4(03), С. 176 - 182
Опубликована: Март 25, 2024
As large language models gain traction in the financial sector, they are revolutionizing workflows of professionals. From data analysis and market forecasting to risk assessment customer management, these demonstrate significant potential value. By automating processing tasks, enhance productivity empower professionals derive deeper insights make more precise decisions. This article explores application conversational intelligent reporting systems, leveraging artificial intelligence models, within industry. It examines how systems streamline processes, facilitate efficient communication, contribute informed decision-making, ultimately reshaping landscape operations.
Язык: Английский
Процитировано
8