Methodology

Steps Followed

1. Analyzing the problem statement & the requirements of the platform

Analyzing the problem in terms of what needs to be to predicted and what kind of observational data is available to make those predictions. Next the method to be used to perform the function must also be finalized.  For this work, a decision tree machine learning algorithm was developed.

2. Collecting and cleaning the data

The next step is to investigate and identify the ideal data available to create the model. The model used here was trained and tested with highly scrutinized and credited datasets. The model was also exposed to questionnaire data and trained to propose accurate and relevant personalization follow-up questions.

3. Preparing data for ML application

Transforming the data into a form that the Machine Learning system can understand.

4. Preparing the User Interface of the model

The user interface is designed for the user to be able to easily interact with the model by giving inputs and the screen displaying the output. There is, firstly, a Home page with the terms of service that describe the guidelines to using the platform. Second, the user is directed to the 'User Profile' page wherein the user can input their personal details and any basic medical conditions. Then, the user is brought to the 'Symptom-based Disease Checker' page. Here are 5 input text boxes which consist of dropdown menu of symptoms and the user can select those one by one. Based on the chosen symptoms, the model creates a panel of questions to ask the user. On pressing the ‘Result’ button, the disease is predicted in the output field. Also, the user can  recieve a list of nearby doctors in the 'Find a doctor' page.

The platform used for the coding was Visual Studio Code along with Scikit-learn (Sklearn) libraries.