Research on "Chatbot for Diabetic Patients" published at the National Conference "Emerging Trends in Intelligent Computing and Communication" , April 13-14, 2012 organised by Department of Information Technology GCET, Greater Noida.
More than Just Lines on a Map: Best Practices for U.S Bike Routes
Research Paper On Chatbot for Diabetic Patient
1. NATIONAL CONFERENCE ON EMERGING TRENDS IN INTELLIGENT
COMPUTING AND COMMUNICATION
APRIL 13-14, 2012
CHATBOT FOR DIABETIC PATIENTS
Gunjan Jain Harshit Gupta Himadri Gupta Himani Singh
CSE Department CSE Department CSE Department CSE Department
IMSEC IMSEC IMSEC IMSEC
Ghaziabad, India Ghaziabad, India Ghaziabad, India Ghaziabad, India
gunjan.jain10@yahoo harshgupta1990@gm himadri.gpt16@gmail singh.himani91@gma
.com ail.com .com il.com
Abstract: Chatbot is a technology in which human can interact
with computers by using natural language(that is
Chatbot is a technology that makes interaction spoken by human). There are a number of chatbot
between human and machines possible by using nowadays that act as a website guidance, online
natural language. In this paper, we proposed an shopping and there are also a chatbots which just
architectural design of a chatbot that will act as a reply to whatever human talks to them. However,
virtual diabetic doctor and make possible for patients some chatbots are also designed which are specific to
to interact with virtual diabetic doctor. This chatbot area of knowledge. It was like the search engine
will allow the patients to have a diabetes control and where user enters the words and engine reply
management advice without the need to go to the according to the search parameters. Although there
hospital. In this design, we used knowledge are some techniques which makes the chatbot to
database(KDB) and pattern matching. On the basis of remember the previous conversation topic, but it
pattern matching between database and KDB, chatbot cannot remember the whole conversation. So, we
will give response to the user as a virtual diabetes propose an architectural design that has the ability to
physician. remember the whole conversation flow and whenever
user wants to see the whole chat, he will be able to
Keywords:
see it.
Chatbot,Knowledge Database(KDB), Pattern
Related Work:
Matching.
ELIZA- It was the first chatbot designed by Professor
Introduction:
Joseph Weizenbaum from Massachusetts Institute of
ISBN: 9788175157262
2. NATIONAL CONFERENCE ON EMERGING TRENDS IN INTELLIGENT
COMPUTING AND COMMUNICATION
APRIL 13-14, 2012
Technology (MIT) [2], which makes possible for Three major components in managing diabetes are
human to interact with computer. In this, input monitoring blood glucose level, proper diet as per
sentences are analyzed by decomposition rules which prescribed by doctor and patients/guardians
are triggered by keywords. And responses are made motivation. The first and most important component
on the basis of reassembly rules. Some of the where patients need to continuously monitor their
technical problems with ELIZA were: (1) the blood glucose level [4]. Following that, patients
identification of keywords, (2) the discovery of needs to continuously go to the hospital for regular
minimal context, (3) the choice of appropriate checkup even if they are in secure control level of
transformations, (4) generation of responses in the their disease. So,we propose to develop a chatbot
absence of keywords, and (5) the provision of an that will function as virtual diabetes doctor which
ending capacity of ELIZA “scripts”. will perform a basic diagnosis on diabetic patients.
A.L.I.C.E.- It is an acronym for Artificial Linguistic The process flows as a regular chatting conversation
Internet Chat Entity, which is a type of dialog-driven using natural language with chatbot which will be a
chatterbot developed in 1995 by Richard Wallace [3]. question and answer session. This session will
ALICE’s knowledge about conversation patterns is continue until patient is successfully diagonised and
stored in AIML files. AIML(Artificial Intelligence get most suitable control advice for their diabetes
Markup Language) is a derivative of XML. AIML condition. In order to do a diagnosis, chatbot will ask
describes a class of data objects and also describes several sequence questions and those questions will
the partial behavior of computer programs. ALICE be selected based on the answers given by patient.
seeks to mimic the conversation rather than
understand it. The algorithm used in programming is as follows:
when a patient says something to a doctor (patient
VPBot- Other than ELIZA and ALICE, there is string) , then that patient string is break into words
which is stored in an array. After that the array
another SQL-based chatbot called VPbot which are
matches every words from the keywords of
used for medical applications. It stores the language knowledge database and if the keywords are found
rules in a relational data model. The VPbot algorithm same then the corresponding response is given to the
accepts three input parameters: vpid, the current topic patient.By using these knowledgable patterns, we
and aa sentence. The vpid is an unique identifier. The make a decision tree about the diseases to get the
topic is an optional parameter, used for handling final decision. If the result comes positive that is if it
pronouns. The output of VPbot is a new sentence is found that the patient has diabetes then chatbot will
provide the patient with cure and treatment and
and a new topic. As stated by Dr. Webber, there are
advice the patient according to knowledge base
some limitations on using single SQL statement patterns. If chatbot finds that patient has no diabetic
because true recursion is not possible [1]. symptoms then it will inform that you are alright and
provide with simple medications so that patient will
Proposed Work: be fit and fine. The diagrammatic representation of
algorithm is shown in figure.1.
Diabetes disease yet cannot be cured by nowadays,
but it can be properly managed and controlled in Conclusion:
order for patient to get a healthy and active life.
ISBN: 9788175157262
3. NATIONAL CONFERENCE ON EMERGING TRENDS IN INTELLIGENT
COMPUTING AND COMMUNICATION
APRIL 13-14, 2012
In this chatbot, we designed the conversation which
is controlled by computer rather than by human by
making the user remain to the conversation topic and
they are not allowed to enter any irrelevant input or
data and if they do, chatbot will response that the
input was not recognizable and keep repeating the
previous question until the keyword is find. The
patient will also be suggested in order to correctly
answers the questions. This design will allow chatbot
to response to the whole conversation as it
specifically designed to be a virtual diabetes
physician.
Acknowledgement:
We would like to thank Dr. Pankaj Agarwal, Head of
Department, CSE department, IMS Engineering
College and Ms. Shruti Keshari, Assistant Professor,
CSE Department, IMS Engineering College for their
support in the process of research.
Figure#1:The Diagramatic Representation of
Algorithm
References:
ISBN: 9788175157262
4. NATIONAL CONFERENCE ON EMERGING TRENDS IN INTELLIGENT
COMPUTING AND COMMUNICATION
APRIL 13-14, 2012
[1] Abbas Saliimi Lokman , Jasmi Mohamad Zain, Questioning’’,Artificial Intelligence Lab ,The
Fakulti Sistem Komputer & Kejuruteraan University of Arizona ,Tucson,Arizona
Perisian,”Designing a chatbot for Diabetic 85721,USA@eller.arizona.edu.
Patients”,university Malaysia Pahang , Lebuhraya tun
Razak 26300 Kuautan,Pahang. [4] Jasni M. Zain and Abdul R.M. Fauzi,”Expectation
and Feasibility of a computer Aided Education in
[2] Joseph Weizenbaum “ELIZA-A Computer Diabetes Urban Area in Malaysia: views from
program for the study of natural language Patients,Healthcare Staff and Hospital
communication between man & machine” Administrators”,in proceedings of the International
Massachusetts Institute Of Technology,Mass Conference of Education,Research and
Communication of the ACM Innovation,ICERI 2008.
volume9,number1(January1966):36-35.
[5] Rik Crutzen,Gjalt-John Y-Peters,Sarah Dias
[3] Robert P. Schumaker, Hsiuchun Chen Portugal,Erwin H.Pisser and Jorne J. Grolleman
“Interaction Analysis of the ALICE chatterbot: A “Chatbot Answering Adoloscents Questions About
Two-Study Investigation of Dialog & Domain Sex,Drugs And Alcohol”,published in 2011.
ISBN: 9788175157262