International Journal of Computer Science Trends and Technology (IJCST) – Volume 11 Issue 5, Sep-Oct 2023
RESEARCH ARTICLE OPEN ACCESS
Health Care Chat Bot By using NLP, Decision Tree and SVM
Mr U. Bhargav Kumar [1], Smt M. Prashanthi [2], Smt D. Madhuri [3]
[1]
PG Scholar, Department of CSE, JNTUA College of Engineering Anantapur.
[2]
Assistant Professor (Contract), Department of CSE, JNTUA College of Engineering Anantapur.
[3]
Assistant Professor (Contract), Department of CSE, JNTUA College of Engineering Anantapur.
ABSTRACT
Healthcare chatbots are designed to provide medical information, answer health-related questions, and offer
support to patients and healthcare professionals through text-based conversations, enhancing accessibility and
efficiency in healthcare services. The increasing population and birth rates in India and reduced death rates due
to medical advancements have created a severe shortage of doctors, resulting in inadequate treatment and patient
fatalities. To address these challenges, the development of an intelligent chatbot that offers guidance to doctors
and patients is crucial, potentially saving lives. Chatbots, versatile virtual assistants, can identify symptoms,
offer diagnoses, and recommend suitable doctors, making them a valuable addition to Healthcare services. The
existing system for a Healthcare chatbot incorporates rule-based and keyword matching, While rule-based and
keyword-matching algorithms offer efficiency and consistency in Healthcare chatbots, it's important to note that
they may have limitations in handling complex or context-dependent queries The proposed system for a
Healthcare chatbot integrates Natural Language Processing (NLP) techniques, Decision Trees, and Support
The concept of Health brought together technologies
Vector Machines (SVM) to create an advanced and intelligent Healthcare chatbot solution.
like the Internet of Health Things, medical Cyber-
Keywords - Chatbot, Natural Language Processing (NLP) techniques, DecisionSystems,
Physical Trees, and
bigSupport Vector Machines
data analytics, (SVM).
and machine
learning, transforming the healthcare industry. Its
primary objective was to provide cost-effective and
I. INTRODUCTION enhanced healthcare services to patients, concurrently
Health professionals face resource constraints and revolutionizing the healthcare business model.
cannot provide constant personal monitoring and Within this context, this paper not only categorized
assistance to patients in their daily routines [1]. To potential healthcare applications but also introduced a
improve healthcare technology electronic patient service-oriented middleware framework designed to
records, and medical devices, it is crucial to minimize facilitate application development within this
errors, especially those that go unnoticed, as paradigm.[4]
clinicians cannot effectively mitigate the harm caused
Healthcare systems have evolved into a market for
by such unnoticed errors [2]. In the context of
companies developing chatbot-based computer
healthcare, machine learning techniques are
programs for patients and doctors [5]. Healthcare
harnessed to leverage the abundant health data
systems that actively engage with aged
generated by IoT. This introduction sets the stage for
people/patients to gather information, monitor health
our exploration of machine learning applications in
conditions, and provide support, especially in case of
medical imaging, natural language processing of
post-hospital situations or in the home environment.
medical documents, and genetic information. We
focused on how machine learning is transforming In the healthcare sector, the adoption of medical
healthcare by aiding in diagnosis, detection, and chatbots is on the rise, primarily due to the expanding
prediction. We also address challenges in data population in India and the shortage of available
utilization, providing insights into the current state of doctors to cater to this rapidly growing populace.
machine learning in healthcare.[3] Even healthcare professionals can make diagnostic
errors, posing risks to patients' lives, as illustrated by
the case of Mohammed Benaziza, a renowned
bodybuilder in the 1990s, who tragically died due to
a misdiagnosis of Hypokalaemia.
To address such issues, medical chatbots play a
crucial role by providing guidance to healthcare
ISSN: 2347-8578 www.ijcstjournal.org Page 4
International Journal of Computer Science Trends and Technology (IJCST) – Volume 11 Issue 5, Sep-Oct 2023
professionals in critical situations and offering initial sleep coach, thus contributing to cost savings and
treatment advice to the general public during user-friendliness in Health Care. [5]
emergencies. The chatbot can also assess specific
ailments through a series of questions and provide Jahnvi Gupta, Vinay Singh, and Ish Kumara[6]
information on recommended precautions and address the common reliance on hospitals for
remedies. Healthcare analysis, diagnosis, and prescriptions. It
introduces a novel approach utilizing the RASA
Additionally, medical chatbots contribute to the
framework to build a chatbot capable of diagnosing
efficiency of the Healthcare system by streamlining
diseases based on user symptoms. Their system aims
administrative tasks like appointment scheduling and
to provide quick and hassle-free answers to health-
medical record management. This efficiency allows
related queries while also recording users' daily
Healthcare providers more time to focus on patient
nutrient intake, offering a potentially transformative
care. Furthermore, as technology continues to
alternative to traditional Healthcare practices.
advance, medical chatbots are poised to expand their
capabilities, potentially revolutionizing the way
In the paper [7] the author explores the evolution of
Healthcare is delivered and accessed, benefiting both
eHealth, driven by technological advancements such
patients and Healthcare providers.
as smartphones and electronic wearable devices. It
II. LITERATURE REVIEW introduces the concept of mHealth and its role in
aiding healthcare through smartphones and IoT
In the paper [1] author explained the faced by health
devices. Additionally, the paper discusses the
professionals who often lack the resources to
increasing importance of information technology in
personally monitor and support patients in their daily
healthcare, leading to the emergence of telemedicine,
lives. They investigate the potential of text-based
which facilitates remote communication between
healthcare chatbots (THCB) to address these
patients and doctors via multimedia. The paper also
challenges, particularly in therapeutic settings beyond
highlights existing telemedicine applications and
traditional on-site consultations. The paper also
proposes a comprehensive approach to eHealth.
introduces an open-source THCB system and its
application in a childhood obesity intervention,
Ming-Chin Yang, Chin-Yuan Huang, and Chin-Yu
showcasing preliminary results indicating promising
Huang [8] address the global rise in obesity,
intervention adherence and scalability, along with
emphasizing the role of mobile applications in public
high patient satisfaction. [1]
health interventions. It introduces the "Smart
Wireless Interactive Healthcare System" (Switches)
Harold Thimble [2] highlights the imperative need to
as a solution for objective data collection and
improve healthcare technologies, including electronic
transmission in real-time. Switches leverage AI-
patient records and medical devices, by reducing use
powered health chatbots to engage users and provide
errors and unnoticed errors that may lead to patient
accurate medical advice, presenting an overview of
harm. The authors emphasize the importance of
their development and implementation.
implementing safety scoring systems to enhance
design safety and inform all stakeholders, including
In the paper [9] the author underscores the
regulators, clinicians, and patients, to put pressure on
importance of Healthcare accessibility and the
manufacturers for safer designs. [2]
potential of AI-driven medical chatbots to reduce
In [5] the author highlights the growing role of
costs and improve medical knowledge availability. It
chatbots in Health Care, describing their potential to
introduces medical chatbots as versatile tools, serving
interact energetically with patients, especially in post-
as medical reference books and offering personalized
hospital or home environments. These chatbots
advice for a healthier life. The paper emphasizes the
bridge the gap between patients and Healthcare
benefits of text-to-text diagnosis bots in engaging
providers, aided by AI and machine learning. The
patients and providing personalized diagnoses based
proposed system aims to assist patients in disease
on symptoms.
identification, acting as both a diagnostic tool and a
ISSN: 2347-8578 www.ijcstjournal.org Page 5
International Journal of Computer Science Trends and Technology (IJCST) – Volume 11 Issue 5, Sep-Oct 2023
Nishargo Nigar and Mohammed Nazim Uddin[10]
highlight the critical role of proper observation in
children's Health Care, particularly in addressing
concerns like childhood obesity and malnutrition. It
introduces the concept of IoT-enabled smart nutrition
cards, tracking various aspects of a child's health and
activity. These cards leverage smartphone sensors,
AI, and recommendation methods, with potential
applications in urgent hospitalization and disease
management.
In the paper [11] the author Ahmed Fadhil focuses on
addressing the rising burden of chronic
noncommunicable diseases by introducing CoachAI,
a conversational agent-assisted health coaching
system. It presents an approach to develop the
chatbot dialogue and coaching platform, aiming to Fig [1] Working of the keyword matching
support physicians in managing patient care and algorithm
reducing their workload. The paper also discusses
patient profiling and physical activity
recommendations, contributing to preventive Health
Care. [11]
III. EXISTING MODEL
In the past, previous systems predominantly relied on Fig [2] Matching case of the keyword and function
either rule-based methodologies or keyword-
In contrast, the rule-based approach operated by
matching algorithms, both of which had limited
making decisions or taking actions according to
capabilities in comprehending natural language
predefined rules or conditions, often drawing on
queries.
human expertise or domain knowledge. This method
Keyword matching, a straightforward algorithm, assessed input data against the established rules,
involves searching for specific words or phrases implementing actions or decisions accordingly. It
within a text and then executing actions or responses could be more intricate than simple keyword
based on predefined keywords or phrases. This matching and typically incorporated logical
approach found common usage in tasks such as operators, conditions, and multiple interacting rules,
content filtering, topic identification, or sentiment finding applications in expert systems, decision
analysis. support systems, and chatbots. Furthermore, the
system allowed for multiple authorizations to access
the chatbot, enhancing its adaptability and user-
friendliness.
ISSN: 2347-8578 www.ijcstjournal.org Page 6
International Journal of Computer Science Trends and Technology (IJCST) – Volume 11 Issue 5, Sep-Oct 2023
Fig [3] working of the rule-based algorithm
Limitations:
• Complex user interface. Making users difficult to
recognize the operations
• Late response to user due to highly dumped data set
• More processing time to process huge data
IV. PROPOSED MODEL
In the proposed Healthcare chatbot system, a
combination of Natural Language Processing (NLP)
techniques, Decision Trees, and Support Vector
Machines (SVM) is employed to create an advanced
and intelligent Healthcare chatbot solution. Fig[4] Working of the NLP chatbot
Natural Language Processing (NLP): NLP is a field Support Vector Machines (SVM): SVMs are used
encompassing linguistics, computer science, and in Healthcare chatbots for classification tasks. They
artificial intelligence. It focuses on teaching involve data collection, feature extraction, data
computers to understand and analyze human splitting, model training, hyperparameter tuning,
language, enabling them to extract information and prediction, and response generation. SVMs find
insights from text documents and facilitate human- decision boundaries to separate different intent
like conversations. Chatbots, as a form of NLP classes, making them suitable for tasks like
software, simulate conversations with users through appointment scheduling or symptom analysis.
natural language, making them a significant Decision Trees: Decision trees are employed for
advancement in human-machine interaction. tasks like symptom analysis, where chatbots need to
Text Pre-processing: Text pre-processing is a make decisions based on a series of conditions. They
crucial step in NLP, involving tasks like tokenization, involve data collection, feature selection, model
lowercasing, removing stop words, stemming, and building, defining stopping criteria, training,
handling punctuation and special characters. It prediction, and response generation. Decision trees
prepares raw textual data for advanced NLP tasks are useful for providing medical advice or
such as sentiment analysis and topic modeling. recommending further actions based on patient
symptoms.
Named Entity Recognition (NER): NER is an NLP
technique that identifies and classifies named entities
in text, including names of people, organizations,
locations, dates, and more. It is essential for
applications like information extraction and
improving search engines' understanding of user
queries.
Intent Recognition: Intent recognition is vital in
conversational AI systems like chatbots, as it helps
determine the user's purpose behind their input.
Machine learning algorithms, including neural
networks, are often used to classify user input into
Fig [5] Decision Tree algorithm working
predefined categories or intents, allowing chatbots to
offer relevant responses. Advantages:
• more visually pleasing
ISSN: 2347-8578 www.ijcstjournal.org Page 7
International Journal of Computer Science Trends and Technology (IJCST) – Volume 11 Issue 5, Sep-Oct 2023
• Less response time to user
• Less processing time
V. EXPERIMENTAL RESULTS
Fig [9] Login page
The above Fig [9] refers login page that offers
convenient access to your account with just your
email address and password. If you're new, simply
click "Register." Need assistance? Explore our
Fig [6] Home page 1 "Help" section for support, check out our "FAQs" for
common questions, or return to the "Home" page for
updates and information
Take Query:
Fig [7] Home page 2
The above Fig [6] and Fig [7] refer to the home page
features a user-friendly interface with options to sign
up, access our AI-powered chatbot for inquiries, and
connect with our support team via the contact section
Registration Page:
Fig [10] Query page
The above Fig [10] refers to the "Take Query" page
functions like a chatbot for health-related queries. It
allows users to input their health symptoms or
Fig [8] Registration page concerns, offering a conversational and user-friendly
The above Fig [8] refers registration page is designed way to seek information and guidance regarding their
to collect essential information for your account well-being.
setup. It includes fields for your first name, last Provide Results:
name, email address, password, phone number, and
address to ensure a seamless and secure user
experience
Login Page:
ISSN: 2347-8578 www.ijcstjournal.org Page 8
International Journal of Computer Science Trends and Technology (IJCST) – Volume 11 Issue 5, Sep-Oct 2023
medical facilities. This makes healthcare more
convenient and accessible to users.
REFERENCES
[1] Tobias Kowatsch, "Text-based
HealthcareChatbots Supporting Patient and Health",
,https://doi.org/10.3929/ethz-b-000218776, 01
October 2017.
[2] Harold Thimbleby, "Improving safety in medical
devices and systems" , College of Science University
of Swansea,
http://dx.doi.org/10.5772/intechopen.92297, Wales
2013.
[3]James P. Brody, Christoper Toh, "Applications of
Machine Learning in Health Care," University of
California, Irvine,
http://dx.doi.org/10.5772/intechopen.92297, January
Fig [11] Results page 2021.
The above Fig [11] refers to the "Provide Result"
page operates as an interactive health chatbot. Users
can input their health symptoms or concerns, and the [4] JAMEELA AL-JAROODI, Nader Mohamed, ,
chatbot will provide valuable information along with EMAN ABUKHOUSA", Health 4.0: On the Way to
recommended precautions, offering personalized Realizing the Healthcareof the Future," Modul
University Dubai, Dubai, United Arab Emirates,
guidance for maintaining well-being.
https://doi.org/10.1109%2FACCESS.2020.3038858,
October 2020.
VI. CONCLUSION AND FUTURE
SCOPE [5] B. Priyadarshini, P. Ponmurugan, , P. Preetha, V.
Preethikadevi, “Healthcare Assisting Chat bot for
Chatbots are an emerging technology with significant symptoms and dosage prediction using IoT", Sri
potential for the future, gaining popularity in various Krishna College of Technology, Coimbatore,
industries. Their impact is expected to be long- Tamilnadu, India, 2018.
lasting, making it exciting to witness the
advancements in this field as they overcome previous [6] Jahnvi Gupta, Vinay Singh, Ish Kumar,
limitations. The development of this system is a "Florence- A Healthcare Chat bot," Galgotias College
response to the growing population in our country. of Engineering & Technology, Greater Noida,
While such systems are already in use in other INDIA,
countries, they are not widely accessible here. The https://doi.org/10.1109/ICACCS51430.2021.9442006
shortage of doctors, especially in government , 2021.
hospitals in our city, is a well-recognized issue.
Introducing a medical chatbot aims to address this [7] Adwitiya Mukhopadhyay, Sreenadh M, Anoop A,
gap by offering medical assistance when doctors are "eHealth Applications: A Comprehensive Approach,"
unavailable, thereby improving the efficiency of the Mysuru Amrita Vishwa Vidyapeetham, India,
healthcare industry and potentially reducing mortality https://doi.org/10.1109/ic-ETITE47903.2020.165,
rates. 2020.
Looking ahead, user-friendly Healthcare chatbots [8] Chin-Yuan Huang, Ming-Chin Yang, Chin-Yu
Huang, "A Chatbot-supported Smart Wireless
have a promising future. They can facilitate efficient
Interactive Healthcare System for Weight Control
doctor consultations, streamline appointment booking and Health Promotion," Proceeding of the IEEE,
processes, and provide map support for locating http://dx.doi.org/10.1109/IEEM.2018.8607399, April
2018.
ISSN: 2347-8578 www.ijcstjournal.org Page 9
International Journal of Computer Science Trends and Technology (IJCST) – Volume 11 Issue 5, Sep-Oct 2023
[9] Ishwarya, A. Divya, Indumathi, Priyasankari, "A
Self Diagnosis Medical Chat bot Using Artificial
Intelligence," proceeding MAT Journal, October
2017.
[10] Nishargo Nigar, Mohammed Nazim Uddin "An
Internet of Things Enabled Intelligent System and
Smart Nutrition Card to Enhance Children’s Health
Consciousness," East Delta University Chittagong,
Bangladesh, June 2018.
[11] Ahmed Fadhil, "Text-based Chatbot Assisted
Health Coaching System: Preliminary Evaluation &
Results," University of Trento, Trento
http://dx.doi.org/10.13140/RG.2.2.22231.32164, May
2019.
ISSN: 2347-8578 www.ijcstjournal.org Page 10
