In an effort to address the persistent challenges in the Colombian healthcare system, I have developed an intelligent chatbot as part of the DocMe project. This chatbot aims to improve accessibility and efficiency in the healthcare system through the use of artificial intelligence.
Context
The Colombian healthcare system faces significant challenges in managing medical appointments, including saturated phone lines and difficulty accessing timely medical services. DocMe aims to address these difficulties through innovative technological solutions, such as the intelligent chatbot, which facilitates interaction and provides quick and accurate responses.
Key Features of the Chatbot
The chatbot in DocMe is designed to handle various situations and medical queries:
Greeting and Welcome
- Input Patterns: Include common greetings such as "hello", "how are you?", among others.
- Responses: The chatbot responds in a friendly and personalized manner, initiating the interaction with the user warmly.
Assistant Information
- Input Patterns: Questions about the chatbot´s name and function.
- Responses: Describes its role as a virtual assistant in managing medical appointments and facilitating the user´s healthcare experience.
Emotional Support
- Input Patterns: Expressions of emotional distress such as "I feel bad" or "I am dealing with a lot of anxiety".
- Responses: Offers emotional support and, if necessary, guides the user to seek professional help, such as counselors or therapists.
Physical Symptoms
- Input Patterns: Indications of physical discomfort such as "I have a fever" or "I feel sick".
- Responses: Provides basic health recommendations and, if necessary, advises the user to schedule an appointment.
Technologies Used
The development of the chatbot is based on the following technologies:
- Python: Used as the main language for backend development and the integration of different project components.
- Flask: Framework used for building the web application´s backend, providing routes and handling HTTP requests.
- Flask-SQLAlchemy for database management: Used to interact with the MySQL database, efficiently managing user and medical appointment data.
- Numpy for mathematical operations: Employed to perform essential numerical calculations and mathematical operations in data processing.
- NLTK for natural language processing: Integrated for analyzing and processing user query text, enabling functions such as tokenization and lemmatization.
- Keras for neural network modeling: Used to develop and train neural network models, improving accuracy in classifying user query intents.
- TensorFlow as backend for Keras: Used as the backend to run deep learning models developed with Keras, ensuring optimized performance.
- waitress as WSGI server: Used as the web server to serve the Flask application in production, providing a lightweight and efficient deployment.
- PyMySQL for MySQL connection: Used to establish and manage the connection between the Flask application and the MySQL database, ensuring data integrity and efficiency.
- SciKit-Learn for machine learning techniques: Employed to implement machine learning algorithms that optimize data classification and pattern detection in user queries.
- PySpellChecker for spell checking: Integrated to improve text processing accuracy by correcting spelling errors in user queries before analysis.
Results and Benefits
The integration of the chatbot in DocMe has significantly improved accessibility and efficiency in managing medical queries. Users can receive quick and accurate responses, reducing administrative burden and improving overall patient satisfaction.
Conclusions
The chatbot in DocMe represents a significant advancement in the digitization of healthcare services, offering an innovative solution to enhance patient-service interaction. This project highlights my ability to apply advanced technologies to solve complex problems, preparing me to face new challenges and contribute to the field of digital health with effective, user-oriented solutions.