| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 4.31 MB | Adobe PDF |
Authors
Abstract(s)
O uso simultâneo de medicina complementar e alternativa com medicamentos convencionais pode levar a interações prejudiciais entre ervas e fármacos, podendo até causar fatalidades em alguns casos. Portanto, é crucial melhorar a compreensão e a consciencialização sobre o uso de ervas e suplementos permitindo, assim, aos profissionais de saúde fornecer orientações adequadas aos pacientes. Os sistemas de suporte à decisão são ferramentas poderosas que podem ajudar os farmacêuticos a tomar decisões diagnósticas e de tratamento. Neste trabalho, um sistema de suporte à decisão foi implementado para identificar interações entre ervas e fármacos ajudando, assim, os farmacêuticos a converter dados científicos em conhecimento aplicável a pacientes. Por outro lado, a cibersegurança tornou-se cada vez mais importante na área da saúde, à medida que a tecnologia é adotada para melhorar o atendimento aos pacientes. Uma área de preocupação prende-se com a autenticação do utilizador em sistemas de informação. A autenticação multi-fator é aplicada para aumentar a segurança e a proteção contínua de dispositivos eletrónicos e serviços críticos. Os dados biométricos, como os padrões de escrita, desempenham um papel fundamental na autenticação, exigindo que os utilizadores comprovem a sua identidade através destas características únicas e individuais. Este trabalho avaliou a aplicabilidade de um novo conjunto de dados de padrões de escrita no desenvolvimento de um sistema de autenticação biométrico inteligente com o objetivo de proteger serviços de saúde. Envolveu duas abordagens de classificação, nomeadamente multi-classe e binária, aplicando três algoritmos de aprendizagem máquina (k-Nearest Neighbors, Random Forest e LightGBM). O algoritmo LightGBM apresenta o melhor desempenho na tarefa de classificação binária, minimizando a rejeição de utilizadores genuínos enquanto evita a aceitação de potenciais atacantes. Este equilíbrio é essencial para assegurar um processo de autenticação eficientes e uma boa experiência ao utilizador. No geral, o trabalho serve como base de referência para investigações futuras, abrindo caminho para soluções avançadas de autenticação biométrica.
The simultaneous use of complementary and alternative medicine with conventional medications can lead to harmful herb-drug interactions and even fatality in some cases. Therefore, it is crucial to improve the understanding and awareness of herbal and supplement usage. This will enable healthcare professionals to provide appropriate guidance to patients and empower users to access reliable information about these therapies. Decision support systems are powerful tools that can assist clinicians in making diagnostic and treatment decisions. In this work, a decision support system was implemented to identify herb-drug interactions and help pharmacists convert scientific data into actionable knowledge for individual patients. On the other hand, cybersecurity has become increasingly important in healthcare as digital technologies are adopted to enhance patient care. One area of concern is user authentication in computer systems. Multi-factor authentication is employed to enhance safety and ongoing protection for computing devices and critical services. Biometrics, such as keystroke dynamics, play a key role in this authentication method by requiring users to prove their identity using multiple factors. The study evaluates the suitability of a novel keystroke dynamics dataset for an intelligent biometric authentication system that safeguards healthcare services. The benchmark performed involves two classification approaches namely, multiclass, and binary. Three algorithms (k Nearest Neighbors, Random Forest, and LightGBM) are used for the analysis. The LightGBM algorithm performs the best in the binary classification task, achieving the lowest false rejection rate and equal error rate values, minimizing the rejection of genuine users while avoiding the acceptance of fraudulent ones. This balance is essential for efficient authentication processes and for ensuring a seamless user experience. Overall, the work serves as a benchmark basis for future investigations, paving the way for advanced biometric authentication solutions.
The simultaneous use of complementary and alternative medicine with conventional medications can lead to harmful herb-drug interactions and even fatality in some cases. Therefore, it is crucial to improve the understanding and awareness of herbal and supplement usage. This will enable healthcare professionals to provide appropriate guidance to patients and empower users to access reliable information about these therapies. Decision support systems are powerful tools that can assist clinicians in making diagnostic and treatment decisions. In this work, a decision support system was implemented to identify herb-drug interactions and help pharmacists convert scientific data into actionable knowledge for individual patients. On the other hand, cybersecurity has become increasingly important in healthcare as digital technologies are adopted to enhance patient care. One area of concern is user authentication in computer systems. Multi-factor authentication is employed to enhance safety and ongoing protection for computing devices and critical services. Biometrics, such as keystroke dynamics, play a key role in this authentication method by requiring users to prove their identity using multiple factors. The study evaluates the suitability of a novel keystroke dynamics dataset for an intelligent biometric authentication system that safeguards healthcare services. The benchmark performed involves two classification approaches namely, multiclass, and binary. Three algorithms (k Nearest Neighbors, Random Forest, and LightGBM) are used for the analysis. The LightGBM algorithm performs the best in the binary classification task, achieving the lowest false rejection rate and equal error rate values, minimizing the rejection of genuine users while avoiding the acceptance of fraudulent ones. This balance is essential for efficient authentication processes and for ensuring a seamless user experience. Overall, the work serves as a benchmark basis for future investigations, paving the way for advanced biometric authentication solutions.
Description
Keywords
Interações entre Ervas e Medicamentos Sistemas de Suporte à Decisão Inteligência Artificial Sistemas Baseados em Regras Aprendizagem Máquina Cuidados de Saúde Herb-Drug Interactions Decision Support Systems Artificial Intelligence Rule-Based Systems Machine learning Healthcare