Virtual Health Assistant using AI

Virtual Health Assistant using AI

(1) Rathod Vijes, (2) Patil Sayali, (3) Landge Shivani, (4) Dhage Aarti, (5) Bombale Girisha

(1) Student, (2) Student, (3) Student, (4) Student, (5) Professor

(1,2,3,4,5) Department of Computer Engineering,

(1,2,3,4,5) SND College of Engineering And Research Centre ,Yeola, Nashik, Maharashtra, India 423401

Abstract – The Virtual Health Assistant using Artificial Intelligence (AI) is an intelligent system designed to provide basic healthcare support and guidance to users through a digital platform. The main objective of this project is to assist individuals in identifying possible health conditions based on their symptoms and to offer preliminary advice, reducing the need for immediate physical consultation in non-critical situations.This system utilizes Natural Language Processing (NLP) and Machine Learning (ML) algorithms to interact with users in a conversational manner, understand their queries, and analyze symptoms. Based on the input, the system predicts potential diseases and suggests precautions, medications, or lifestyle changes. It can also provide information about nearby healthcare services, emergency assistance, and general health awarenessThe Virtual Health Assistant improves accessibility to healthcare, especially for people in remote areas, and saves time by offering instant responses. It is designed to be user-friendly, cost-effective, and available 24/7. However, it is not intended to replace professional medical advice but to act as a supportive tool for early diagnosis and health management.

Keywords – Artificial Intelligence, Virtual Health Assistant, Machine Learning, Natural Language Processing, Healthcare Technology, Disease Prediction

I. INTRODUCTION

Today, technology is changing the way we take care of our health, and Artificial Intelligence (AI) is playing a big role in this transformation. Tools like Machine Learning (ML) and Natural Language Processing (NLP) help computers understand human language and learn from data. Because of this, it has become possible to build smart systems that can answer health-related questions and guide people in a simple and quick way.

A Virtual Health Assistant is one such system that acts like a digital helper for basic healthcare needs. It can talk with users, understand their symptoms, and suggest possible health conditions along with simple precautions. This not only saves time but also helps people get instant support without visiting a hospital for minor issues.This kind of technology is especially useful for people who live in areas where medical facilities are not easily available. It provides support anytime, making healthcare more accessible and convenient. While it cannot replace a real doctor, it is very helpful for early guidance and spreading awareness about health. Overall, the

Virtual Health Assistant shows how AI can make healthcare easier, faster, and more user-friendly for everyone.

• AI Processing (Analysis Stage)

  In the analysis stage, the system processes the users input using Artificial Intelligence (AI) techniques to understand and interpret the given symptoms. When a user enters their health-related query, the system first applies Natural Language Processing (NLP) to break down the text, identify important keywords, and extract relevant symptoms.After extracting the symptoms, the system converts the text data into a structured format using techniques like tokenization and vectorization. This structured data is then passed to the Machine Learning (ML) model, which has been trained on medical datasets. The model analyzes the input by comparing it with learned patterns and relationships between symptoms and diseases.

  Based on this analysis, the system predicts possible health conditions and calculates the probability or likelihood of each condition. It may also consider factors such as symptom severity and combinations to improve accuracy.Overall, the AI processing stage acts as the core of the system, where raw user input is transformed into meaningful insights, enabling accurate predictions and effective health recommendations.

• Output

  The system provides accurate and instant responses to user health queries based on the symptoms entered. It analyzes the input using Machine Learning and Natural Language Processing to predict possible diseases and suggest basic precautions or remedies.

  Additionally, the Virtual Health Assistant generates useful health insights such as recommended actions, lifestyle tips, and when to seek medical attention. It ensures users receive quick guidance without delay.Overall, the output of the system is a user-friendly and reliable health support solution that improves accessibility, saves time, and assists in early-stage health decision-making.

• Overall Impact and Significance

  The system provides accurate and instant responses to user health queries based on the symptoms entered. It analyzes the

  input using Machine Learning and Natural Language Processing to predict possible diseases and suggest basic precautions or remedies.

  Additionally, the Virtual Health Assistant generates useful health insights such as recommended actions, lifestyle tips, and when to seek medical attention. It ensures users receive quick guidance without delay.Overall, the output of the system is a user-friendly and reliable health support solution that improves accessibility, saves time, and assists in early-stage health decision-making.

  • METHODOLOGY

The development of the Virtual Health Assistant using Artificial Intelligence follows a systematic and well-defined methodology to ensure accuracy, reliability, and user satisfaction. The entire process is divided into multiple stages, starting from data collection to system deployment and evaluation.

1. DataCollection:

   The first step involves gathering relevant healthcare datasets from reliable sources such as medical databases, research papers, and publicly available health records. The dataset includes information about symptoms, diseases, causes, and possible treatments. High-quality data is essential because the accuracy of the system depends on it.

2. DataPreprocessing:

   Once the data is collected, it is cleaned and prepared for analysis. This step includes removing duplicate entries, handling missing values, correcting inconsistencies, and converting the data into a structured format. Text data related to symptoms is also normalized to improve understanding by the system.

3. FeatureExtractionandSelection:

   Important features such as symptoms, severity levels, and patient inputs are identified and selected. These features are used to train the model effectively. Techniques like tokenization and vectorization are applied to convert textual data into numerical form for processing.

4. NaturalLanguageProcessing(NLP):

   NLP techniques are used to enable the system to understand and interpret user queries. When a user enters symptoms in natural language, the system processes the input, identifies keywords, and extracts meaningful information. This allows smooth and human-like interaction between the user and the system.

5. ModelDevelopment(MachineLearning):

   In this phase, Machine Learning algorithms such as Decision Tree, Naïve Bayes, or Support Vector Machine (SVM) are used to train the model. The model learns the relationship between symptoms and iseases based on the training data. Multiple models can be tested, and the best-performing one is selected based on accuracy and performance.

   • Information Extraction and Ontology Modeling

     In the Virtual Health Assistant system, information extraction is used to identify and collect important details from user input and medical datasets. When a user enters symptoms in natural language, the system applies Natural Language Processing (NLP) techniques to extract key information such as symptoms, duration, and severity. This step helps convert unstructured text into meaningful and structured data that the system can understand and process.

   • Similarity Computation and Ranki Mechanism

     In the Virtual Health Assistant system, the similarity computation stage is used to compare the users input symptoms with existing medical data. After extracting key symptoms, the system converts them into numerical form and matches them with stored symptom-disease patterns in the dataset. Techniques such as cosine similarity or distance-based methods are used to measure how closely the users symptoms match known cases.Once similarity scores are calculated, the system applies a ranking mechanism to prioritize the most relevant results. Diseases with higher similarity scores are ranked at the top, indicating a higher probability of occurrence. This helps the system present the most likely health conditions first, along with appropriate suggestions.

   • System Implementation

     The system implementation phase focuses on building and integrating all components of the Virtual Health Assistant into a working application. The system is developed using programming languages such as Python, along with libraries and frameworks that support Machine Learning (ML) and Natural Language Processing (NLP), such as TensorFlow, Scikit-learn, and NLTK.The implementation begins with the development of the backend, where the trained ML model is integrated. This backend handles tasks such as symptom analysis, disease prediction, and response generation. The NLP module is also implemented here to process user input and extract relevant information.The system also includes a database to store medical data, user queries, and model outputs. This helps in improving system performance and enables future enhancements. APIs

     The system implementation describes how the proposed virtual health assistant is designed, developed, and deployed using AI technologies. It integrates multiple components such as data processing, machine learning models, and user interfaces to deliver accurate and real-time healthcare support.

     may be used to connect the frontend with the backend for smooth data flow.

   • Evaluation Metrics

     Evaluation metrics are used to measure how well the Virtual Health Assistant performs in predicting diseases and providing accurate responses. These metrics help ensure that the system is reliable and effective.The most common metric is accuracy, which shows how many predictions made by the system are correct out of the total predictions. However, accuracy alone is not enough, especially in healthcare applications.Therefore, additional metrics such as precision, recall, and F1-score are used. Precision measures how many of the predicted diseases are actually correct, while recall measures how many actual disease cases are correctly identified by the system. The F1-score provides a balance between precision and recall, giving an overall performance measure.

   • Overall Outcome

     The Virtual Health Assistant successfully provides a smart and efficient solution for basic healthcare support using Artificial Intelligence (AI). The system is able to understand user inputs, analyze symptoms, and predict possible diseases with good accuracy using Machine Learning (ML) and Natural Language Processing (NLP) techniques.It delivers quick and reliable health guidance, including precautions, lifestyle suggestions, and recommendations on whether to seek medical attention. This helps users make informed decisions about their health without immediate hospital visits for minor issues.The project also improves accessibility and convenience, especially for people in remote areas, by offering 24/7 support through a user-friendly interface.

   • Proposed Framework

     The proposed framework for the Virtual Health Assistant is designed as a multi-layered architecture that ensures smooth data flow, accurate prediction, and meaningful user interaction. It integrates Artificial Intelligence (AI) techniques such as Machine Learning (ML) and Natural Language Processing (NLP) to provide an intelligent healthcare support system.

     The framework begins with the User Interaction Layer, where users input their symptoms or health-related queries through a web or mobile interface. This input is then passed to the Data Preprocessing Layer, where the text is cleaned, normalized, and prepared for analysis.Next, the Information Extraction Layer uses NLP techniques to identify key

     symptoms and relevant medical information from the user input. This structured data is then processed in the AI Processing (Analysis) Layer, where ML models analyze the symptoms and compare them with trained datasets.

     The processed data is then handled by the Similarity Computation and Ranking Layer, which finds the closest matches between user symptoms and known disease patterns. Based on this, the Predictive Analytics and Recommendation Layer generates possible disease predictions along with precautions and health advice.An Ethical and Fairness Layer is integrated to ensure data privacy, reduce bias, and maintain transparency in the system. Finally, the results are presented back to the user through the interface in a clear and understandable format.

     Overall, this framework provides a structured and efficient approach for delivering accurate, fast, and user-friendly healthcare assistance using AI technologies.This proposed framework provides a complete and structured approach for developing a Virtual Health Assistant. By combining NLP, ML, data retrieval, and user interaction, the system delivers accurate, fast, and accessible healthcare support. It ensures efficient processing of user input while maintaining reliability, scalability, and user trust.

     • ALGORITHMS USE

       Decision Tree Classifies diseases by splitting symptoms into decision rules.

       Naïve Bayes Classifier Predicts disease using probability based on given symptoms.

       Random Forest Combines multiple decision trees to improve prediction accuracy.

       K-Nearest Neighbors (KNN) Finds similar symptom cases and predicts based on nearest data.

       Support Vector Machine (SVM) Separates diseases using an optimal boundary for accurate classification.

       Natural Language Processing (NLP) Understands and processes user input in human language.

     • System Architecture

A. The Virtual Health Assistant using AI is an intelligent, real-time healthcare support system designed to provide users with instant medical assistance and personalized health guidance. The system utilizes Artificial Intelligence (AI) and Natural Language Processing (NLP) to understand user queries, analyze symptoms, and generate accurate health-related responses in a conversational format.

The assistant functions as a virtual healthcare companion, enabling users to interact through both text and voice interfaces. By processing user input, the system identifies symptoms, interprets intent, and predicts possible health conditions using trained machine learning models. Based on this analysis, it provides relevant suggestions such as preventive measures, basic treatment advice, and lifestyle recommendations.

RESULTS AND DISCUSSION

The developed Virtual Health Assistant system was tested sing a variety of symptom inputs to evaluate its performance and effectiveness. The system successfully analyzed user queries and provided relevant predictions along with appropriate health suggestions. The results indicate that the integration of Machine Learning (ML) and Natural Language Processing (NLP) enables accurate understanding of user input and efficient disease prediction.

During testing, the system demonstrated good accuracy and consistency in identifying common diseases based on symptoms. The use of similarity computation and ranking mechanisms helped in prioritizing the most relevant results, improving the overall quality of predictions. Additionally, the chatbot interface provided quick responses, making the interaction smooth and user-friendly.From a performance perspective, the system showed fast response time, which is important for real-time healthcare assistance. Users were able to receive instant feedback without delays, enhancing usability and convenience. The recommendation layer also provided useful guidance such as precautions and suggestions for further medical consultation.

• System Performance and Efficiency

  The performance of the Virtual Health Assistant is evaluated based on its accuracy, speed, and ability to handle user queries effectively. The system demonstrates good performance in understanding user input using Natural Language Processing (NLP) and predicting possible diseases through Machine Learning (ML) models. It provides relevant and meaningful responses for common symptoms, ensuring reliable preliminary guidance.In terms of efficiency, the system processes user input quickly and generates responses in real time. The optimized data preprocessing and lightweight ML models help reduce computation time, resulting in fast response generation. This makes the system suitable for continuous user interaction without noticeable delays.

  The chatbot architecture also ensures smooth conversation flow through effective dialog management, allowing the system to handle multiple queries efficiently. The chatbot architecture also ensures smooth conversation flow through effective dialog management, allowing the system to handle multiple queries efficiently. Additionally, the use of structured datasets and a knowledge base improves the speed of information retrieval and enhances overall system performance.

  However, system efficiency may vary depending on factors such as dataset size, model complexity, and server capacity. Regular updates and optimization can further improve accuracy and response time.Overall, the Virtual Health Assistant achieves a good balance between performance and efficiency by providing accurate predictions, fast responses, and a smooth user experience, making it a practical solution for basic healthcare assistance.

• Accuracy and Matching Quality

  The accuracy of the Virtual Health Assistant reflects how correctly the system predicts diseases based on user-provided symptoms. By using Machine Learning (ML) models trained on medical datasets, the system achieves good accuracy for common diseases where symptom patterns are well-defined. The use of Natural Language Processing (NLP) further improves accuracy by correctly interpreting user input and extracting relevant symptoms.

  The matching quality refers to how well the system compares user symptoms with existing disease profiles. Through techniques like similarity computation and pattern matching, the system identifies the closest matches between input symptoms and stored data. Higher similarity scores indicate better matches, which are then ranked to present the most relevant results to the user.The ranking mechanism ensures that the most probable diseases appear at the top, improving the usefulness of the output. This combination of accurate prediction and effective matching enhances user trust and decision-making.

• Fairness and Bias Mitigation

  Fairness and bias mitigation are essential to ensure that the Virtual Health Assistant provides accurate and equal healthcare guidance to all users. Since Machine Learning (ML) models learn from data, any imbalance or limitation in the training dataset can lead to biased predictions. To address this, the system is trained on diverse and representative datasets that include different age groups, genders, and health conditions, reducing the risk of unfair outcomes.

  The system also applies techniques to detect and minimize bias during model development. This includes analyzing model outputs across different user groups and ensuring that no group is consistently receiving less accurate or misleading results. Regular evaluation and testing help identify any hidden biases in predictions.In addition, the framework maintains transparency by clearly communicating that the system provides general health guidance and not definitive medical diagnoses. This helps prevent over-reliance on AI decisions.

• Comparative Evaluation

  The proposed Virtual Health Assistant is evaluated by comparing it with traditional healthcare systems and basic rule-based chatbot systems. This comparison highlights the advantages and improvements achieved through the use of Artificial Intelligence (AI), Machine Learning (ML), and Natural Language Processing (NLP).

  Compared to traditional healthcare methods, where users must visit hospitals or clinics for even minor issues, the Virtual Health Assistant provides instant responses and basic health guidance. This reduces time, cost, and effort, especially for non-critical situations. It also improves accessibility for users in remote areas where medical facilities may not be easily available.When compared to rule-based chatbots, which follow predefined rules and limited responses, the proposed system performs better due to its learning capability. ML-based models can analyze patterns in data and provide more accurate and flexible predictions. NLP enables the system to understand natural language, making interactions smoother and more user-friendly than rigid keyword-based systems.

• Discussion and Insights

  The development and evaluation of the Virtual Health Assistant highlight the growing potential of Artificial Intelligence (AI) in transforming basic healthcare services. By integrating Machine Learning (ML) and Natural Language Processing (NLP), the system is able to understand user queries, analyze symptoms, and provide meaningful health guidance in real time. This demonstrates how AI can simplify initial diagnosis and improve access to healthcare support.

  One of the key insights from the project is the importance of data quality. The accuracy of predictions largely depends on the dataset used for training. Well-structured and diverse medical data leads to better performance, while limited or biased data can reduce reliability. This emphasizes the need for continuous data updates and improvement.Another important observation is the role of user interaction and experience. A simple and conversational interface significantly enhances usability, making it easier for users to communicate their problems. The chatbot-based approach makes the system more engaging and accessible, even for non-technical users.The project also highlights the balance between automation and responsibility. While the system provides quick and helpful suggestions, it cannot replace professional medical advice. This reinforces the need for clear communication about the systems limitations and proper ethical considerations.

• Summary of Findings

1. Improved Accessibility to Healthcare InformationThe system shows that users can easily access basic medical guidance anytime without visiting a hospital. It reduces dependency on immediate doctor availability for minor queries and symptoms.

2. Efficient Symptom AnalysisUsing AI-based analysis, the assistan is able to process user-input symptoms and map them to possible health conditions. This helps in early awareness and encourages timely medical consultation when needed.

3. Fast and Accurate Response SystemThe project findings highlight that the system significantly reduces response time compared to traditional consultation methods. The recommendation engine provides relevant suggestions quickly based on stored medical knowledge.

4. Personalized Health Recommendations

   The assistant is capable of giving personalized advice such as diet suggestions, precautionary measures, and basic medication guidance depending on user input and health history (if included).

5. Data Preprocessing Improves AccuracyProper preprocessing of medical data (like cleaning, normalization, and feature extraction) improves the accuracy of prediction and ensures better matching between symptoms and possible diseases.

   • Overall Outcome

     The overall outcome of the Virtual Health Assistant using AI project is the successful development of an intelligent system that provides basic healthcare support through automated symptom analysis and recommendation generation.

     The system effectively demonstrates how artificial intelligence can be applied in the healthcare domain to assist users in understanding possible health conditions based on their symptoms. It improves accessibility by offering instant responses without requiring direct interaction with medical professionals for minor concerns.The assistant is able to process user inputs, extract relevant medical information, and generate meaningful suggestions such as possible diseases, precautions, and general health advice. This helps users take early preventive action and encourages timely consultation with doctors when required.

     The project also shows that integrating data preprocessing, machine learning techniques, and a recommendation engine enhances the accuracy and relevance of results. Additionally, the inclusion of an ethical layer ensures that the system provides safe, non-harmful guidance and clearly states its limitations.Overall, the system proves to be a useful decision-support tool that reduces the burden on healthcare services, improves health awareness, and provides a foundat

     • CONCLUSION

       The Virtual Health Assistant using AI project successfully demonstrates the practical application of artificial intelligence in the healthcare domain to provide intelligent, accessible, and user-friendly medical support. The primary objective of the system was to develop a virtual platform that can assist users in understanding their health conditions based on symptoms and provide appropriate preliminary guidance. This objective has been effectively achieved through the integration of AI techniques such as data preprocessing, information extraction, similarity matching, and predictive analysis.

       The system shows that AI can significantly enhance the way healthcare information is delivered to users. Instead of relying only on traditional methods of consultation, users can quickly interact with the assistant to get possible disease predictions, health recommendations, and preventive measures. This reduces delays in understanding health issues and promotes early awareness, which is crucial for better health outcomes.

       One of the key achievements of this project is the development of a structured pipeline that processes user input, extracts meaningful medical features, and compares them with an existing medical knowledge base. The recommendation mechanism ensures that the responses are relevant and useful. Additionally, the system is designed with an ethical and safety-focused approach, ensuring that it does not provide direct medical prescriptions but instead acts as a supportive advisory tool encouraging users to consult healthcare professionals when necessary.

       The project also highlights the importance of data quality and preprocessing in improving system performance. Clean, well-structured data improves prediction accuracy and ensures more reliable outputs. Furthermore, the inclusion of fairness and bias considerations helps in making the system more dependable across different user groups.

       From a usability perspective, the virtual assistant provides a simple and interactive interface, making it accessible even to non-technical users. This enhances user engagement and ensures that the system can be used in real-world scenarios without difficulty.However, the project also acknowledges certain limitations. The system may not always handle rare or complex diseases with high accuracy due to limited datasets. It cannot replace professional medical diagnosis and is intended only as a supportive tool. Continuous improvement and updates are necessary to keep the system aligned with evolving medical knowledge.

       In conclusion, the Virtual Health Assistant using AI represents a meaningful step toward the integration of artificial intelligence in healthcare support systems. It demonstrates how technology can be used to improve accessibility, awareness, and decision-making in health-related situations. With further enhancements, such systems have strong potential to become valuable tools in modern healthcare ecosystems, assisting both users and professionals in delivering better health outcomes.

     • REFERENCES

1. G. Zhao, Z. Li, X. Wang, W. Ning, X. Zhuang, J. Wang, Q. Chen, Z. Mo, B. Chen, and H. Chen, Herb Community Detection from TCM Prescription Based on Graph Embedding, in 2018 7th International Conference on Digital Home (ICDH), IEEE, 2018, pp. 318323. doi: 10.1109/ICDH.2018.00062.

2. W. Li, The Confinement-Escape Problem of a Defender Against an Evader Escaping from a Circular Region, IEEE Transactions on

   Cybernetics, vol. 46, no. 4, pp. 10281039, 2016. doi: 10.1109/TCYB.2015.2503285.

3. I. Ullah, N. U. Amin, A. Almogren, M. A. Khan, M. I. Uddin, and Q. Hua, A Lightweight and Secured Certificate-Based Proxy Signcryption (CB-PS) Scheme for E-Prescription Systems, IEEE Access, vol. 8, pp. 199197199212, 2020. doi: 10.1109/ACCESS.2020.3033758.

4. Artificial Intelligence in Intelligent Traffic Signal Control, IEEE Access, 2025.

5. J. Miles and N. Qamar, Med Conformity: Enhance Adherence with Prescription Opioids, in 2018 IEEE International Conference on Healthcare Informatics (ICHI), IEEE, 2018, pp. 340341. doi: 10.1109/ICHI.2018.00048.

6. M. A. Sadikin and S. U. Sunaringtyas, Implementing digital signature for the secure electronic prescription using QR-code based on Android smartphone, in 2016 International Seminar on Application for Technology of Information and Communication (ISemantic), IEEE, 2016, pp. 306311. doi: 10.1109/ISEMANTIC.2016.7873856.

7. P. Srivastava and N. Singh, Automatized Medical Chatbot (Medibot), in 2020 International Conference on Power Electronics & IoT Applications in Renewable Energy and its Control (PARC), IEEE, 2020,

   pp. 351354. doi: 10.1109/PARC49193.2020.236624.

8. J. Gupta, V. Singh, and I. Kumar, Florence A Health Care Chatbot, in 2021 7th International Conference on Advanced Computing and Communication Systems (ICACCS), IEEE, 2021, pp. 504508. doi: 10.1109/ICACCS51430.2021.9442006.

9. W. Astuti et al., Predicting Frequently Asked Questions (FAQs) on the COVID-19 Chatbot using the DIET Classifier, in East Indonesia Conference on Computer and Information Technology (EIConCIT), IEEE, 2021. doi: 10.1109/EICONCIT50028.2021.9431913.

10. N. Rosruen and T. Samanchuen, Chatbot Utilization for Medical Consultant System, in 2018 3rd Technoogy Innovation Management and Engineering Science International Conference (TIMES-iCON), IEEE, 2018, pp. 15. doi: 10.1109/TIMES-ICON.2018.8621678.

1. W. Astuti, D. P. I. Putri, A. P. Wibawa, Y. Salim, Purnawansyah, and A. Ghosh, Predicting Frequently Asked Questions (FAQs) on the COVID-19 Chatbot using the DIET Classifier, in East Indonesia Conference on Computer and Information Technology (EIConCIT), IEEE, 2021. doi: 10.1109/EICONCIT50028.2021.9431913.

2. J. Gupta, V. Singh, and I. Kumar, Florence A Health Care Chatbot, in 2021 7th International Conference on Advanced Computing and Communication Systems (ICACCS), IEEE, 2021, pp. 504508. doi: 10.1109/ICACCS51430.2021.9442006.

1. J. Miles and N. Qamar, Med Conformity: Enhance Adherence with Prescription Opioids, in 2018 IEEE International Conference on Healthcare Informatics (ICHI), IEEE, 2018, pp. 340341. doi: 10.1109/ICHI.2018.00048.

2. M. A. Sadikin and S. U. Sunaringtyas, Implementing digital signature for secure electronic prescription using QR-code based on Android smartphone, in 2016 International Seminar on Application for Technology of Information and Communication (ISemantic), IEEE, 2016, pp. 306311. doi: 10.1109/ISEMANTIC.2016.7873856.

3. N. Rosruen and T. Samanchuen, Chatbot Utilization for Medical Consultant System, in 2018 3rd Technology Innovation Management and Engineering Science International Conference (TIMES-iCON), IEEE, 2018, pp. 15. doi: 10.1109/TIMES-ICON.2018.8621678.

4. P. Sharma et al., Optimization of Traffic Signal Phases Using Deep Q-Learning and Genetic Algorithms, International Journal of Traffic Management and Control, vol. 25, no. 4, pp. 183197, 2023.

5. J. Miles and N. Qamar, Med Conformity: Enhance Adherence with Prescription Opioids, in 2018 IEEE International Conference on Healthcare Informatics (ICHI), IEEE, 2018, pp. 340341. doi: 10.1109/ICHI.2018.00048.

6. G. Zhao, Z. Li, X. Wang, W. Ning, X. Zhuang, J. Wang, Q. Chen, Z. Mo, B. Chen, and H. Chen, Herb Community Detection from TCM Prescription Based on Graph Embedding, in 2018 7th International Conference on Digital Home (ICDH), IEEE, 2018, pp. 318323. doi: 10.1109/ICDH.2018.00062.