Healthcare Chatbot Using Decision Tree Algorithm
Healthcare Chatbot Using Decision Tree Algorithm
Volume 7, Issue 10, October – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Healthcare Chatbot using Decision Tree Algorithm
Dr. Renuka R Devi Himanshu Banerjee, Baibhab Saha
Designation: Associate Professor SRMIST,
Department: C. Tech SRMIST, Chennai Chennai
Abstract:- Artificial Intelligence(AI) and Machine The chatbot is simply a code that runs on a software by
Learning(ML) has become an integral part of our day to copying human conversation and infer a fruitful result.The
day life. A large number of technologies are coming forth user will be given queries and they will answer based on
to help us and make our life easier. Chatbot or virtual which symptoms will be diagnosed and a disease or health
assistant is a technology that is coming forward in this dilemma will be understood. This will help the patient to refer
regard. Previously, we had simple questions and answers to the specific specialist. For instance, the patient can refer to
or FAQs in the name of assistant, but now the chatbot will a cardiologist if a heart problem is diagnosed by the
diagnose the disease based on queries answered by the chatbot.The design of the chatbot will suit the user and they
users. can get a quick and accurate diagnosis based on the
symptoms they produce in the questionnaire.
The model will be trained by using a highly
diversified and accurate database with symptoms and II. LITERATURE SURVEY
diseases.The proposed model uses a decision tree
algorithm to initiate a top down follow up and identify the For our project we took into account several other
complexity of the patient and produce a outcome.The user previously successful works on the chatbots from the
is provided a question and answer approach and a healthcare domain.This paper focuses on our model which
preferable and highly accurate diagnosis is produced. resorts to question and answer patterns with the user.The user
will feel that they are having a dialogue with another person
Keywords:- Healthcare Chatbot, Artificial Intelligence, and not a chatbot.
Machine Learning,Decision Tree Algorithm.
The diagnosis done by our chatbot will be more accurate
I. INTRODUCTION than the existing models as our dataset is more enriched and
factual. Our dataset gives accurate conclusions and better
We live in the information age, the business with more outputs with proper diagnosis.
information survives the market. AI plays a vital role in
streamlining the data into the computer which gives a useful The chatbot uses a decision tree classifier as the
result to make human life easier. AI stands for Artificial algorithm and gives proper diagnosis.The algorithm is quick
Intelligence, the science which allows machines to improve and gives better conclusions.
capabilities of average human intelligence. In today’s world,
self-driving cars have come into existence. This shows the
immense possibilities in development of machines which
would help humans in large scales. At this stage AI is the way
to explore the limits of technology. Over the last decade,
people and organizations have gained trust on AI.
Fig. 1: Trustworthiness of AI
The above chart shows that 59% of the population agrees Table 1: References from previous papers
with the fact that AI has been helping the human world in
various ways, while 24% of the population did not have the
idea about the roles of AI, the other 17% of the
population were clueless.
IJISRT22OCT969 www.ijisrt.com 1604
Volume 7, Issue 10, October – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
III. PROPOSED WORK It is best to implement the learning algorithm using
Python libraries such as Scikit learn, matplotlib, numpy,
The model we proposed has an interactive chat window pandas, etc. We clean the dataset and the exploratory data
that connects with the patient and gives them the feel as if they analysis is done using the mentioned libraries. After the data
are talking to a person. cleaning and analysis module, we move into the training
phase of the learning model using the resultant data.
The chatbot provides a question and answer pattern to the
user and asks them about the symptoms, intensity of the The Decision Tree classifier is used to train the model.
symptoms, number of days they are facing those symptoms and A Decision tree classifier is a Supervised learning
finally the chatbot gives a diagnosis. algorithm. It can be used to tackle both classification and
regression
The user gets the idea about which specialist they should
visit and the severity of their situation can also be understood
by them. Our idea is to build a dataset that is very accurate and
very close to the actual medical diagnosis performed by a
doctor. The chatbot would be very user friendly and easy to
use,it would have a very attractive interface.
problems, however it is more preferable for classification
problems.
IV. IMPLEMENTATION
The plan to implement the HealthCare chatbot is to use
Python based libraries. The chatbot will be implemented as a
Python application. Fig. 3: Flowchart of Decision Tree
Python can be used to build GUI [Graphical User In a Decision Tree classifier which is a tree-
Interfaces] using the Tk GUI library. Tkinter acts as an structured classifier, the interior nodes represent the
interface to the Tk GUI library. Tkinter, an optimal GUI features of the used dataset. The branches in the tree
library in Python can be used efficiently to add GUI structure are the rules to make decisions and classify them
components in an easy way. It facilitates powerful object- into predictions and the leaf nodes are the generated
oriented interconnectivityto the Tk GUI library. outcomes. The Decision Tree is a graphical representation to
determine all the possible outcomes to a problem based on
In order to be helpful to the users, the chatbot will be fed the training dataset. To form a tree structure, CART
with an enriched symptom mapped disease dataset which Classification and Regression tree algorithm is used. The
would help the learning algorithm to give appropriate results tree formation is implemented by asking questions and
based on the symptoms a user query carries. based on the answer(Yes/No), the splitting off of the
branches is done.
Decision Tree imitates human decision making, and
therefore is pretty easy to understand.
Additionally we have another feature which would
direct the patient to a specific doctor based on their disease.
The input to the chatbot system will be mapped in the
trainingdataset and accordingly the disease is predicted.
V. RESULTS DISCUSSION
The finished model of the healthcare chatbot is a
functional system with a large dataset of human health
problems. The chatbot gives a very accurate diagnosis of
the current health problem of the user as per the given user
query.The product produces a very accurate answer to the
health problems in a very short period of time.The user gets
a clear picture of his/her health scenario and get an idea of
which specialist doctor to refer going forward.The machine
learning algorithm used provides a very wholesome review
on our dataset and helps diagnose the health situation of the
Fig. 2: Flowchart of proposed implementation
user based on their reply to the predefined questionnaire
about their symptoms.Hence, the resultant chatbot has a
great scope going forward and achieving a pinnacle in the
medicalfield.
IJISRT22OCT969 www.ijisrt.com 1605
Volume 7, Issue 10, October – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
VI. CONCLUSION
In the medical field, a chatbot could help panicked
patients to get a specific prediction of the disease that they may
have and a proper roadmap that would ensure their proper
recovery. This project is a way to help the medical world by
being an immediate repellant to the patients in emergency
situations. Gradually we can enhance the dataset and do further
study in order to increase intelligence of the system. From the
survey taken in the population across the globe we found that
trust on the artificial intelligence and machine learning
algorithms has increased over the years. The chatbot has a
huge role in the future in the medical and other surrounding
fields,this also has a massive role in revolutionizing the way
medical diagnosis as compared to the methods humans used
previously.
REFERENCES
[1.] IJCRT [International Journal of Creative Research
Thoughts] - October, 2021:
https://ijcrt.org/papers/IJCRTH020011.pdf
[2.] International Journal of Innovative Research in
Computer Science and Technology- May, 2018:
https://www.ijircst.org/DOC/2_irp620.pdf
[3.] International Journal of Innovative Research in
Engineering & Management (IJIREM)
https://www.ijirem.org/DOC/1-A-Survey-on-Healthcare-
Chatbot-Using-Machine-Learning.pdf
[4.] ITB Journal of Information and Communication
Technology
https://www.researchgate.net/publication/326658271
[5.] Journal of Emerging Technologies and Innovative
Research https://jetir.org/papers/JETIR2204602.pdf
