Abstract

Health is the most important in every human’s life. Weekly or monthly check up of
one’s health is most important for the prevention and also to stay healthy. Nowadays,
the individual is not having that much time to go for health check-up. Recently, due to
covid-19, no one are willing to go to hospital for health check-up due to the fear of
spreading virus. In this situation, technology plays and important role. In this project, we
have used Machine Learning. Machine Learning is the study of computer algorithms
that improve automatically from the previous experience. It is widely used nowadays
and it is the most efficient domain in health care. We will develop a GUI to get the
symptoms from the user. The models used in this project are Naive Bayes and Decision
Tree. The output is the disease, the accuracy of model, its definition and the treatment
of the particular disease based on the symptoms given by the individual. As we all know
the saying which tells that “Prevention of the disease at an early stage is much better
than the cure which we take after we get affected by the disease”. This project shows
detailed explanation of how to find the diseases from symptoms, so that the individual
can contact the respective doctor and stay healthy at an early stage.

Methodology

Project contains three parts:
3.8.1 DATASET COLLECTION.
3.8.2 TRAIN AND TEST THE MODEL.
3.8.3 DEPLOY THE MODELS.
 Dataset Collection- We had collected dataset from kaggle notebooks.
The dataset contains the symptoms and the corresponding disease. It
contains 4920 rows and 133 columns.
 Train and Test the model- We had used the Naïve Bayes Classifier as
a model to train the dataset. After training, we had tested the model and
found its accuracy.
 Deploy the models- Deployed Naïve bayes by creating interface to get
the name, symptoms of an individual. By this, we will get the disease
and accuracy of model as the output. We have also created a chatbot
using Decision Tree which helps an individual to get the corresponding
disease by checking whether he/she is being faced by the symptoms.
Following are the steps to do this project (use Jupyter Notebook):

1. Collect the dataset.
2. Import the necessary libraries.
3. Visualise the dataset.
4. Train the dataset using Naïve Bayes classifier and Decision Tree.
5. Test the model and find the accuracies of both.
6. Deploy the model- a) Naive Bayes as GUI Interface using Tkinter
   b) Chatbot using Decision Tree
7. Predict the disease based on the symptoms given by the user.

Results

Doctors and medical professionals are always required in case of an
emergency. In the current situation of COVID-19, where essential resources are
unavailable and people are also not willing to go outside in fear of spreading virus,
our prediction system will be very helpful for finding the disease based on the
symptoms in the early stage and get the correct diagnosis of a disease.This also
helps in reduction of the cost and give the correct and fast result.
4.1. WORKING
In this way machine learning when implemented in healthcare can help in
satisfying the individual and also take care of their particular disease easily. Naive
Bayes, which is the most easy model helps to get the idea about the disease of an
individual based on the symptoms he/she have, and get the treatment easily by
contacting the concern doctor. We have also created chatbot using decision tree
by which an individual can easily find the disease by chatting. Accuracy is defined
as the ratio of sum of TP and TN to the sum of TP, TN, FP, FN.
Algorithm
Accuracy
Decision Tree
97%
Naive Bayes
100%