Implementation of a Chatbot System using AI and

Implementation of a Chatbot System using

AI and NLP
Tarun Lalwani, Shashank Bhalotia, Ashish Pal, Shreya Bisen, Vasundhara Rathod

Abstract— For using software applications, user domain where it is operating. The most popular example
interfaces that can be used includes command line, today is the Amazon’s Alexa. Chat bots are at almost every
graphical user interface (GUI), menu driven, form- place, one can see it at every second website they visit. A bot
based, natural language, etc. The mainstream user is helpful in answering queries related to information which
interfaces include GUI and web-based, but occasionally might be unreachable at that website easily. Most of the
the need for an alternative user interface arises. A websites avail users with chat bots to aid them to go through
chatbot based conversational user interface fits into this what the websites facilitate. They are turning out to be our
space. The chatbot is a class of bots that have existed in virtual assistants in everyday lives.
the chat platforms. The user can interact with them via
graphical interfaces or widgets, and the trend is in this A. Basics of chat bot
direction. They generally provide a stateful service i.e. A chatbot is an artificially intelligent creature which can
the application saves data of each session. On a college’s converse with humans. This could be text-based, or a spoken
website, one often doesn’t know where to search for conversation (in case of voice-based queries). Chat bots are
some kind of information. It becomes difficult to extract basically used for information acquisition. It can run on the
information for a person who is not a student or local PCs and mobile phones, though most of the time
employee there. The solution to these comes up with a it is accessed through the internet. It can be compelling,
college inquiry chat bot, a fast, standard and captivating and spell-bounding. It is a conversational agent
informative widget to enhance college website’s user which interacts with users in a certain domain or on a
experience and provide effective information to the particular topic with input in natural language sentences.
user. Chat bots are an intelligent system being Mainly a chatbot works by a user asking some question or
developed using artificial intelligence (AI) and natural initiating a new topic of discussion. Chat bots can be referred
language processing (NLP) algorithms. It has an as software agents that pretend as human entity. These are
effective user interface and answers the queries related the agents with AI embedded and using NLP they can answer
to examination cell, admission, academics, users’ to user questions. Predefined knowledge base helps develop a
attendance and grade point average, placement cell and response to the query.
other miscellaneous activities. B. Chatbot for College
Keywords:- Artificial intelligence, chat bot, knowledge The need for college inquiry system arises due to various
base, lemmatization, natural language processing, reasons which include: the slow nature of college website, an
semantic sentence similarity, wordnet. outsider would not know where to search for a particular
piece of information, difficult for the person outside
I. INTRODUCTION college’s domain to extract information. The smart solution
for all the drawbacks lends to the need of the system. The
Nowadays, we see the chat bots everywhere Chat bots are college inquiry system will provide the response by
the source of answers to the users questions in any summarizing the query and then output answers, it also
particular domain where it is operating. Chat bots are the provides selective information what the user wants. A
source of answers to the users questions in any particular college system will dispense all answers relating to domains
such as admission, examination cell, notice board,
attendance, placement cell and other miscellaneous domains.
Manuscript Received April 25, 2018. The major features of the chatbot are:
Tarun Lalwani, Shashank Bhalotia, Ashish Pal, Shreya
 College admission related queries could be answered
Bisen, Student, Department of Computer Science and
Engineering, Shri Ramdeobaba College of Engineering and through it.
Management, Ramdeo Tekdi, Gittikhadan, Katol Road, Nagpur,  Viewing user profiles and retrieves attendance and
Maharashtra, India grade/ pointers.
Vasundhara Rathod, Assistant Professor, Department of  College students can get information about
Computer Science and Engineering, Shri Ramdeobaba College of examinations to be held.
Engineering and Management, Ramdeo Tekdi, Gittikhadan, Katol  College students can fetch particulars about placement
Road, Nagpur Maharashtra, India Phone(91)-(712)-2580011 activities.
(email: shrreyabisen@gmail.com)

Copyright © 2018. Innovative Research Publications. All Rights 1

Electronic copy available at:

 International Journal of Innovative Research in Computer Science & Technology (IJIRCST)
ISSN: 2347-5552, Volume-6, Issue-3, May 2018
College inquiry system will act as a fast, standard and that query DOI: 10.21276/ijircst.2018.6.3.2
informative widget to enhance college website’s user
experience and bestow users with righteous information.
The bot will analyze user’s queries and understand users’
message and then reply accordingly. It uses AI & NLP.
This way users’ time and efforts will be saved and s/he
will be equipped with effective answers.
The objectives of this application are:
 To analyze users queries and understand users
message.
 To provide an answer to the query of the user very
effectively.
 To save the time of the user since s/he does not have
to personally go to the college for inquiry.
 This system will help the student to be updated about
the college activities.
 The system will reply using an effective GUI which
implies that as if a real person is talking to the user.

II. PROPOSED SYSTEM

A. Context Identification:
Pre-processing is applied to the input text to standardize
the input as per the system’s requirement. Based on the
keywords used in the text, appropriate context is
recognized.
B. Personal Query Response System (Module-1):
Upon receiving personal queries like CGPA, attendance,
etc., the authenticity of the user is checked through user-id
and password. If the user detail is invalid, an appropriate
response is sent.
If the user authenticates successfully, the input text is
processed to extract keywords. Based on the keywords,
information required by the user is understood and the
information is provided from the database.
C. AIML Response System
(Module-2):
If the user is trying to make a normal conversation with the
bot, the input is mapped to an appropriate pattern in
Artificial Intelligence Modeling Language (AIML) files. If
the response is available, it is sent to the user. Other data
provided to the chatbot such as username, gender, etc. are
also saved. If the pattern is not available in AIML files, a
random response is sent suggesting “Invalid Input”.
D. Query Analysis and Response System (Module-3):
When a user wants some information pertaining to college,
the response will be provided through this module.
If the input matches a pattern in the AIML files, the
appropriate response will be sent to the user. If the AIML
files have no entry for that particular query pattern,
keywords are fetched from the input.
An algorithm to check sentence similarity (NLP) is
applied to the modified input to check its similarity with
the questions of a predefined question-set, whose answers
are available.
If a sentence is retrieved with confidence> 0.5, we
return the answer of that question as the response.
If no questions map to the user input, the input is
saved in a log file for improvement of the system by the
admin. The administrator can incorporate the answer to

Copyright © 2018. Innovative Research Publications. All Rights 2

Electronic copy available at:

 Implementation of a Chatbot System using AI and
in the knowledge base if s/he finds it convenient. Also, a
random response is sent to the user suggesting “Answer
not available”.
E. Context Reset:
Once the user is satisfied with the response of bot and
does not wish to chat further, he/she has the option to log
out of the system or simply exit. Once the user exits the
system, all input parameters are automatically reset.
III. DESIGN
Use Case Diagram:
(User and admin roles)

Fig 1: Use Case Diagram of user and admin

roles

Data Flow Diagram:

Fig 2: Zero level DFD of Chatbot system

Fig 3: First level DFD of Chatbot system

Copyright © 2018. Innovative Research Publications. All Rights 3

Electronic copy available at:

 International Journal of Innovative Research in Computer Science & Technology (IJIRCST)
ISSN: 2347-5552, Volume-6, Issue-3, May 2018
DOI: 10.21276/ijircst.2018.6.3.2

Use Case Diagram: College Related Query Response Activity (Module-3):

(Context identification)

Fig 4: Use Case of context identification

Activity Diagram:
Personal Query Response Activity (Module-1): Fig 7: Activity Diagram of College Related Query Response

IV. IMPLEMENTATION
A. AIML:
To create our knowledge base for normal conversation, we
have used AIML files to store the question and answers pair.
When user converses with our chat bot, the input is matched
to patterns listed in AIML files and corresponding answer is
returned as response.

The sample AIML file structure is as:

<aiml version = "1.0.1" encoding = "UTF-8"?>
<category>
<pattern> HELLO USERNAME</pattern>
<template> Hello User!</template>
</category>
</aiml>
Fig 5: Activity Diagram of Personal Query Response
Activity B. Lemmatization and POS Tagging Using WordNet:
Information extraction from the input text was done by
Normal Conversation Response Activity (Module-2):
extracting keywords. For example, “What is the current
placement scenario?” contain “current”, “placement” and
“scenario” as the keywords. Appropriate Lemmas of the
keywords were found using Lemmatization and POS
tagging, to group together the different inflected form of the
words. For example, requiring, require and required should
map to require. WordNet from Python’s “nltk” package was
used for this purpose.
C. Semantic Sentence Similarity:
There are various combinations in which user can input the
same query. For example,
Q1: What is the notice regarding PG courses re-registration?
Q2: Tell me about re-registration in PG courses in our
Fig 6: Activity Diagram of Normal Conversation Response college.

Copyright © 2018. Innovative Research Publications. All Rights 4

Electronic copy available at:

 Implementation of a Chatbot System using AI and
Q1 and Q2 both mean the same thing (same sense). Also, to provide the proper link. The database storage includes
there will be many more combinations for this same query information about questions, answers, keywords, and logs.
and finding all such combinations will not be feasible. The We have also developed an interface. The interface
scalability and performance of the system will also get developed will have two parts, one for users and the other for
affected. To overcome this problem, similarity is found out the administrator.
between the user input and the queries present in the VII. FUTURE ENHANCEMENTS
available question set (whose answers are available with
the system). The query which matches with the input with Instead of AIML based bot, other algorithms can be
maximum score gets selected (if greater than threshold) implemented. We can include voice-based queries. The users
and the appropriate response is returned. will have to give voice input and the system will give the text
Similarity score for two sentences is calculated by output. Also, after successful execution of chatbot in college
averaging the similarity of the individual keywords of domain, we can implement it in other domains like medical,
those sentences. Each keyword of the first sentence is forensic, sports, etc. It will be beneficial in all the
matched with every keyword of the second sentence to fields as without spending much time, we are accessing the
find the word with maximum similarity to it. Then the relevant information and that too without any sorting.
similarity score of individual words are averaged to
represent the sentence similarity. VIII. ACKNOWLEDGMENT
To find the word similarity- Path Similarity and Wu-
We would like to extend our sincere thanks to Prof.
Palmer (WUP) Similarity is used.
Vasundhara Rathod for giving us good guidelines throughout
Path similarity computes shortest number of edges
numerous consultations.
from one word sense to another word sense, assuming a
hierarchical structure like WordNet. In general, word
senses which have a longer path distance are less similar REFERENCES
than those with a very short path distance, e.g. man, dog [1] Yuhua Li, David McLean, Zuhair A. Bandar, James D.
versus man, tree (expectation is that man is more similar to O’Shea, Keeley Crockett, “Sentence Similarity Based on
dog than it is to tree). The Wu-Palmer metric weights the Semantic Nets and Corpus Statistics”, IEEE Transactions
edges based on distance in the hierarchy. on Knowledge and Data Engineering, Volume 18 - No. 8,
D. Log File: August 2006.
[2] Emanuela Haller, Traian Rebedea, “Designing a Chat-bot
We have maintained a log file which stores the inputs that Simulates an Historical Figure”, IEEE Conference
which the chatbot was not able to answer. Admin can see Publications, July 2013.
the log and add the response of relevant sentences to the [3] Pratik Slave, Vishruta Patil, Vyankatesh Gaikwad, Girish
knowledge base. This would help in improvement of the Wadhwa, “College Enquiry Chat Bot”, International
chatbot knowledge system i.e. the databases. Journal on Recent and Innovation Trends in Computing
and Communication, Volume 5, Issue 3, March 2015.
V. APPLICATIONS [4] “AIML Based Voice Enabled Artificial Intelligent
 It enables the students to be updated with college Chatterbot”, International Journal of u- and e- Service,
activities. Science and Technology Volume 8 - No. 2, 2015.
 It saves time for the students as well as teaching and [5] Amey Tiwari, Rahul Talekar, Prof. S. M. Patil, “College
non-teaching staffs. Information Chatbot System”, International Journal of
 It is providing us a readily available information Engineering Research and General Science, Volume 2,
source without taking any physical efforts. Issue 2, April 2017.
 It is easily accessible and saving time and money also. [6] Rachit Kulkarni, Ankit Methwani, Nakul Pawar, Charmi
Valecha, Pooja Shetty, “College Chat-bot”, International
VI. CONCLUSIONS Journal of Advanced Research in Computer Engineering
& Technology, Volume 6, Issue 4, April 2017.
It is often impossible to get all the data on a single interface [7] Chaitrali S. Kulkarni, Amruta U. Bhavsar, Savita R.
without the complications of going through multiple forms Pingale, Prof. Satish S. Kumbhar, “BANK CHATBOT -
and windows. The college chatbot aims to remove this An Intelligent Assistant System Using NLP and Machine
difficulty by providing a common and user-friendly Learning”, International Research Journal of Engineering
interface to solve queries of college students and teachers. and Technology, Volume 4, Issue 5, May 2017.
The purpose of a chatbot system is to simulate a [8] Yash Mehta, Shreya Sawkar, “The college chatbot”,
human conversation. Its architecture integrates a language International Journal of Computer Applications, Volume
model and computational algorithm to emulate information 173 - No. 7, September 2017.
online communication between a human and a computer [9] Prof. K. Bala, Mukesh Kumar, Sayali Hulawale, Sahil
using natural language. Pandita, “Chat-Bot For College Management System
The college student and employees can freely upload Using A.I”, International Research Journal of
their queries. The chatbot provides fast and efficient search Engineering and Technology, Volume 4, Issue 11, Nov
for answers to the queries and gets the relevant links to 2017.
their question. A background research took place, which
included an overview of the conversation procedure and
tries to find out the relevant keywords related to that query

Copyright © 2018. Innovative Research Publications. All Rights 5

Electronic copy available at:

 International Journal of Innovative Research in Computer Science & Technology (IJIRCST)
ISSN: 2347-5552, Volume-6, Issue-3, May 2018
DOI: 10.21276/ijircst.2018.6.3.2

Author’s Profile

Tarun Lalwani is currently in his last Shreya Bisen is currently in her last
semester pursuing the degree of Bachelor of semester pursuing the degree of Bachelor of
Engineering from Shri Ramdeobaba College Engineering from Shri Ramdeobaba College
of Engineering and Management, Nagpur, of Engineering and Management, Nagpur,
Ramdeo Tekdi, Gittikhadan, Katol Road, Ramdeo Tekdi, Gittikhadan, Katol Road,
Nagpur - 440 013 (M.S.) (India). He is a Nagpur - 440 013 (M.S.) (India). She is a
student at Department of Computer Science student at Department of Computer Science
and Engineering. His academic projects and Engineering. His academic projects include: Customer
include: AVOID Q : Quick Billing System, an innovative Relationship Management(CRM) Software System, as an industry
approach to minimize problems faced by people in billing project; Enhance security for Cloud storage using hybrid
section; Real Time Intrusion Detection System using Machine Cryptography (a way to secure files on clud storage using
Learning, etc. encryption algorithms); Emergency Response Application, which
which help users through an androd app to locate nearby help
centers. She is also a member of the Computer Society of India
(CSI).

Shashank Bhalotia is currently in his last

semester pursuing the degree of Bachelor of
Engineering from Shri Ramdeobaba College Vasundhara Rtahod is currently the
of Engineering and Management, Nagpur, Assistant Professor at Department of
Ramdeo Tekdi, Gittikhadan, Katol Road, Computer Science and Engineering at Shri
Nagpur - 440 013 (M.S.) (India). He is a Ramdeobaba College of Engineering and
student at Department of Computer Science Management, Nagpur, Ramdeo Tekdi,
and Engineering. His academic projects Gittikhadan, Katol Road, Nagpur - 440 013
include: Content Security Management System (to minimize (M.S.) (India). She has 3 years of teaching
piracy of paid content); Gestured Computer Controller experience as of now and 1 year industry
(controlling Computer actions through hand gestures) and experience. 3 international journal research papers have been
Gestured Calculator (a calculator taking all inputs through fingers published in her account. The subjects of her specialization include:
and providing output in text and speech). He was also a part of an Object Oriented Technologies, Computer Networks, Web
open source project called “XCOS-ON-WEB” during his Application Development.
internship (Summer 2017) at IIT Bombay. He is a member of the
Computer Society of India (CSI). His research paper publication
include a research paper titled: “A novel Approach to Compute
Steiner point in Graph: Application for Network Design”
(IJCSE).

Ashish Pal is currently in his last

semester pursuing the degree of Bachelor
of Engineering from Shri Ramdeobaba
College of Engineering and Management,
Nagpur, Ramdeo Tekdi, Gittikhadan,
Katol Road, Nagpur - 440 013 (M.S.)
(India). He is a student at Department of
Computer Science and Engineering.
His academic
projects include: Media content Security System; Dynamic/
Static Hand Gesture Recognition System.

Copyright © 2018. Innovative Research Publications. All Rights 6

Electronic copy available at: