See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/347423315

Smart College Chatbot using ML and Python

Conference Paper · December 2020

DOI: 10.1109/ICSCAN49426.2020.9262426

CITATIONS READS

19 24,136

1 author:

Ashok Kumar K
Matrusri Engineering College
24 PUBLICATIONS 56 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

network on chip View project

Internet of Things View project

All content following this page was uploaded by Ashok Kumar K on 06 February 2021.

The user has requested enhancement of the downloaded file.

 SMART COLLEGE CHATBOT USING ML AND
PYTHON
Hrushikesh Koundinya K Ajay Krishna Palakurthi
Electronics and Communications Engineering Electronics and Communications Engineering
Matrusri Engineering College Matrusri Engineering College
Hyderabad, India Hyderabad, India
hrushikeshkulkarni1999@gmail.com ajaykrishna9800@gmail.com

Vaishnavi Putnala
Electronics and Communications Engineering Dr. Ashok Kumar K
Matrusri Engineering College Electronics and Communications Engineering
Hyderabad, India Matrusri Engineering College
vaishuvv47@gmail.com Hyderabad, India
kashok483@gmail.com

Abstract-The days of solely engaging with a service through make increasingly complex techniques. That, yet these days
a keyboard are over. Users interact with systems more and there are additionally several Natural Language Processing
more through voice assistants and chatbots. A chatbot is a (NLP) [1] and intelligent systems that could comprehend
computer program that can converse with humans using human language. Artificial intelligent systems learn
Artificial Intelligence in messaging platforms. Every time themselves and retrieve insight by perusing required electronic
the chatbot gets input from the user, it saves input and articles that have been existed on the web.
response which helps chatbot with little initial knowledge A chatbot (otherwise called a chatterbox, Bot, or Artificial
to evolve using gathered responses. With increased Conversational Entity) is an AI program [2] that copies human
responses, precision of the chatbot also gets increases. The discussions including content and communication in natural
ultimate goal of this project is to add a chatbot feature and language utilizing artificial intelligence methods, for example,
API for Matrusri Engineering College. This project will Natural Language Processing (NLP), picture and video
investigate how advancements in Artificial Intelligence and processing, and voice analysis. Chatbot for college
Machine Learning technology are being used to improve management system has been created utilizing artificial
many services. Specifically it will look at development of intelligence algorithms that examine the user queries. This
chatbots as a channel for information distribution. The chatbot system is an internet application that gives an answer
program selects the closest matching response from closest to the broken down queries of an user. Users simply need to
matching statement that matches input utilizing WordNet, choose the classification for inquiries and afterward ask the
it then chooses response from known selection of question to the bot that utilizes for noting it. Artificial
statements for that response. This project aimed to intelligence has been incorporated to respond to the user's
implement online chatbot system to assist users who access inquiries. Then the user can procure the fitting solutions to
college website, using tools that expose Artificial their inquiries.
Intelligence methods such as Natural Language Processing, The appropriate responses are given utilizing artificial
allowing users to communicate with college chatbot using intelligence algorithms. Users won't need to go actually to the
natural language input and to train chatbot using college or college website for requests. Users need to enlist to
appropriate Machine Learning methods so it will be able the system and needs to login to the system. After login users
to generate a response. There are numerous applications can get to the different helping pages. There will be different
that are incorporating a human appearance and intending helping pages through which users can chat by asking
to simulate human dialog, yet in most part of the cases questions related with college activities. The system answers
knowledge of chatbot is stored in a database created by a to users' queries with the assistance of effective Graphical
human expert. User Interface (GUI). The user can question about the college
Keywords- Chatbot; Artificial Intelligence; Machine related activities with the assistance of this web application.
learning; WordNet; Natural Language Processing College related activities, for example, admissions, academics,
Intake, and other social activities. It will support the
I. INTRODUCTION undergraduates/other user to be refreshed about the college
The improvements in the fields of inter-networking and activities. A chatbot is an Artificial Intelligence program that
information technology have been intricate in executing an can converse with people in natural language, the manner in
Artificial Intelligent (AI) systems. These systems are drawing which we collaborate with one another. It can trade a human
nearer of human activities, for example, choice emotionally for some undertakings of replying inquiries. A chatbot is a
supportive networks, robotics, natural language processing, specialist that assists users in utilizing natural language. It was
and so forth. Indeed, even in the artificial intelligent fields, worked as an endeavor to trick people. A few uses of chatbots,
there are some hybrid strategies and adaptive techniques that for example, User care, customer support and so on utilizes
Artificial Intelligence Markup Language (AIML) [3] to visit


 
     

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on February 06,2021 at 10:07:14 UTC from IEEE Xplore. Restrictions apply.
with users. One of the foremost objectives of chatbots is to altogether different meaning. The user is permitted to ask any
take after a smart human and entangle the recipient of the number of questions with respect to institution. Chatbots after
discussion to comprehend the genuine working along with receiving query from user checks confidence [6] score and
different designs and abilities for their use has generally gives legitimate response to the user question. The keyword
widened. These chatbots can demonstrate adequate to trick the match calculation is done where the user inquiry went through
user to believe that they are "talking" to an individual, 3 keyword matching algorithm [7]. If this matching of
however, they are limited in improving their insight base at keywords fails then at that point query is sent through 2 and 1
runtime, and have typically next to zero methods for keeping keyword matching with the database. Even then if the query
track of all the discussion information. Chatbots utilize AI to doesn't get the right keyword match, at that point the chatbot
arrive at counterfeit intelligence helping them to comprehend application sends No Answer Found as a reply.
the user question, what's more, give a suitable reaction. The The utilization of logic adapters to choose a response is
chatbots are created utilizing the Artificial Intelligence another algorithm used for chatbot applications. The aim of an
Markup Language (AIML) for imparting or cooperating with input adapter is to get input from bot source, and then convert
the user. This comprises software that will be made up of it into a format that makes chatbot understand. The chatbot
utilizing Artificial Intelligence and will assist the user in system uses a special logic adapter that allows to pick the
chatting with a machine. The user can ask the systems like fitting response from all the responses. The Multi Logic
typically did to other humans. Adapter is used to choose a single response from the responses
The remaining of paper as follows: section-II provides returned by all of the logic adapters that the chat bot has been
literature survey and section-III presents proposed system with configured to use. Preprocessing of information is done by
methodology. Section-IV propounds results and discussion word embedding. Here each word is mapped to a vector and
and finally, section-V concludes the paper. the vector structure is spoken to in one-hot encoded structure
[8] which implies 1 represents the presence of word and 0 for
II. LITERATURE SURVEY everything else. Natural Language ToolKit (NLTK) is a
By utilizing the field of Artificial Intelligence, one can python library which offers assistance for Natural Language
develop numerous applications one of that is mentioned in this Processing (NLP). NLTK [9] has inbuilt tokenizers. The
paper is a college chatbot system. In spite of the fact that NLTK incorporates a wide scope of tokenizers which are as
chatbot can be deployed in various fields like marketing, per the following norm, letters, path, words, keywords, class,
education, banking, clinical and finance. Research is being N-gram, pattern and so on. The most usually utilized tokenizer
done in making the regular rule based chatbots to be is the word-punkt tokenizer [10] which parts the sentences at
informative, responsive and complete the correspondence in a the blank spaces. The precision, speed and effectiveness of the
conversational human language. This requires the NLTK tokenizers is exemplary. Administrator signs in to the
incorporation of Natural Language Processing (NLP) and portal and can perform activities like erase invalid answer or to
Machine Learning (ML) technologies into the college chatbot include explicit answer of a specific inquiry. With the
system. There are various approaches to do as such. Selecting assistance of computerized reasoning, the chatbot application
a fitting technique depends on the area of the chatbot, the answers the question asked by the users.
functionalities it expects to give, the language of
correspondence, the end client, and so forth. Some of the III. PROPOSED SYSTEM
approaches are versed in this literature survey. This College Chatbot System is a web based application
Michael Maudlin created "Chatter Bot Algorithm" in 1994 which gives responses to the user queries. The system
and published in the book Julia and was used to answer the architecture of the chatbot system is shown in the Fig. 1.
queries. Taking this initial idea, further projects were Firstly, Chatbot responds to the user by greeting him/her and
developed to create a chatbot system. The user need to login to then asks user to login into the system by providing his/her
Chat-Bot application. At exactly that point the user is mail. Then the user finds the buttons in the UI which
permitted to submit complaints and queries. When user query corresponds to the different categories of the college. After
is submitted to the bot, context of the query is recognized and going through the buttons the chatbot system asks the user, is
NLP is applied. WordNet calculation [4] and grammatical it helpful in giving the response. If the user is not able to find
forms labeling are utilized to distinguish the feeling of the the required response he/she can continue the chat with the
words. User questions are checked in the knowledge database. college chatbot system by briefly elaborating their queries.
If the appropriate response is discovered, at that point that Then chatbot system applies Machine Learning algorithms to
answer is sent to that user. If a particular query isn't found in the break down the user queries.
the database such inquiries are replied by administrator. When Once the user asks query, the keywords in the query is
the administrator answers the query, at exactly that point the detected using WorldNet Algorithm. As the query description
appropriate response is sent to the user. Question alongside can change from one person to another person. The same
answer is put in database so that at whatever point such query may be asked in a different ways by the users. One user
inquiries will be posed with the intention that they get asks a query so simply and clearly while another user may
addressed legitimately from the database. Because of this request same query in a completely different manner. So it is
administrator doesn't have to address same query physically required to find what is the exact information user seeks to
any longer. Different algorithms such as Porter Stemmer know and to find a correct response for the corresponding user
Algorithm [5] is used for expelling suffixes from words in query. The chatbot system firstly removes the stop words from
English. Word request vector process is used for estimating the user input, if they are present in the queries asked by the
word request closeness between two sentences. Sentences with user. After removing the stop words from the user queries,
precisely same words yet in different order may bring about tokenization and lemmatization [11] process are done.


 
     

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on February 06,2021 at 10:07:14 UTC from IEEE Xplore. Restrictions apply.
 Fig. 2. Flowchart for User Module
Fig. 1. College Chatbot system architecture
the chatbot. User’s query is first checked in database. If the
Tokenization is a process of taking a set of text or text and query is valid then suitable response is given to the user. If the
breaking it up into its individual words or sentences. query is invalid then chatbot requests user to ask queries
Lemmatization is the process of gathering the different regarding the college.
inflected forms of a word so they can be dissected as a solitary d. Providing feedback: After the chat, the chatbot takes
item and is a variation of stemming. From there spell checker feedback from the user. Feedback is taken in order to know the
[12] is used to identify and rectify spelling mistakes present in users experience with the chatbot. If the user gives feedback
the query, then by using the sentence similarity and WordNet positively then the bot thanks the user and provides a box to
Algorithm [13] a suitable response is explored in the enter any further queries. If the user gives feedback negatively
knowledge database [14]. WorldNet is a semantic and lexical then bot asks the user to elaborate his/her query in order to
database for the English language. It is used to group English respond. Username of the user is also stored and helps admin
words into the set of synonyms called synsets [15], it gives to track user actions.
short definitions and utilization models, and records various At the other end, admin who is responsible for
relations among these synonym sets or their individuals. If the maintaining the college chatbot system up to date has several
response is found in the database it is displayed to the user, functions to perform such as add the query to the database,
else the system notifies the admin about missing response in modify the data, delete the data, and view feedback given by
the database and gives a predefined response to the user. user and so on. All the functions to be performed by the admin
Admin can write the missing response into the database by are outlined below in detail as shown in Fig. 3.
logging into the admin block in website so that if the user asks a. Login: System has only one admin (there is no
the same query next time, he/she may get the suitable registration for admin). Admin has to login by providing
response. At the end of conversation the college chatbot his/her username and password entered password is encrypted
system collects the feedback from users to improve the system using SHA-256 Encryption algorithm. The login details are
efficiency. validated against the username and password which are stored
The functions of the user are to ask queries, provide in the database and are encrypted using SHA-1 Encryption
feedback and so on. All the functions to be performed by the algorithm. If the details provided are matching with the
user are outlined below in detail as shown in Fig. 2. database then the admin can get the access of college chatbot
a. Login: After clicking on the chatbot provided in the system.
college website. The chatbot system greets the user and b. Add query: If admin proceeds to add dataset, then the
requests the user to provide the mail id. After which the chatbot allows to add the query in three options that is addition
chatbot starts chatting with the user. of question, addition of answer and selecting the respective
b. Botindex: When the user proceeds to choose chatbot to category into which dataset is added.
get an answer to his/her query, the chatbot displays a page to c. View dataset: If admin proceeds to view dataset, then
select few options regarding college and identifies his/her the chatbot allows to view the dataset category wise. The
category of query. If the user gets his query cleared then the chatbot also gives an additional two options that is delete the
task of chatbot is completed. dataset and modify the dataset.
c. Asking Queries: If the user is not satisfied with rule d. Delete query: If admin proceeds to delete query, then
based response, then the chatbot system requests to enter the chatbot allows to delete the query from view page itself by
his/her query in words and the suitable response is given by selecting respective category.


 
     

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on February 06,2021 at 10:07:14 UTC from IEEE Xplore. Restrictions apply.
 Fig. 3. Flowchart for Admin Module
e. Modify query: If admin proceeds to modify existing
query, then the chatbot allows to modify the query from view
page itself by selecting respective category.
f. Change password: If admin wants to change the
password, then the chatbot allows to change the password. To
change the password, admin must provide old password, new
password and re-enter the new password in the change
password webpage. Thus creating a new password which is
encrypted and stored in the code.
g. Viewing invalid dataset: If admin proceeds to view Fig. 4. Chatbot menu containing display of options and asks user to provide
invalid dataset, then the chatbot allows to view the dataset mail id
category wise. The invalid data is the data which the user has As soon as user chooses a particular category, the chatbot
given negative feedback or the queries for which the chatbot is collects user mail id. If the users query is not solved by
unable to respond. The chatbot also gives an additional two options then chatbot system gives additional dialogue box to
options that is delete and modify corresponding query. write his/her question regarding college. User can ask any
h. Edit Static answers: The text displayed when user number of queries to chatbot system regarding college. Some
selects buttons in GUI of the chatbot system can be updated or sample queries asked by the user are shown in Fig. 5. Chatbot
modified by the admin. The admin can update the information system answers all the queries of users without any delay.
which is obtained by selecting the button in the webpage or
can change the function of the button by rewriting it in the
database.
All the functions permit the administrator to perform any
action through the website without going through the database.

IV. RESULTS AND DISCUSSION

Chatbot system is implemented to meet academic
requirements of the users. Simulation or Generating response
from a chatbot is a knowledge-based one. Wordnet is
responsible for retrieving the responses and in this case, it
contains all logics that is triggered whenever the user context
is matched. When a user begins asking queries in the chatbot
Graphical Use Interface (GUI). The query is searched in the
database. If the response is found in the database it is
displayed to the user else the system notifies the admin about
the missing response in the database and gives a predefined
response to the user.
Admin can write the missing response into the database
by logging into the admin block in website. The chatbot is
based on AIML language which is type of Extensible Markup
Language (XML). This helps the different type of user to get
the information like latest news, university rank holders,
timetables, updates regarding college exams and activities and Fig. 5. Chatbot answering queries of users
other academic information. Some pictures of the proposed After the chat, chatbot system asks user to provide
chatbot system is shown in Fig.4, 5 and 6 respectively. By feedback as shown in Fig. 6. This feedback system is
giving choices users can discover their answers in a single employed to know whether the user is satisfied with the
click.


 
     

Authorized licensed use limited to: Lovely Professional University - Phagwara. Downloaded on February 06,2021 at 10:07:14 UTC from IEEE Xplore. Restrictions apply.
 V. CONCLUSION
In this project we made a college specific chatbot system that
can be custom fitted to education domain chatbot, the addition
of this chatbot system in the college website will make the
webpage more user interactive as it responds to the user
queries very accurately as it is a domain specific chatbot
system, and furthermore we had investigated our college
chatbot system design stages and a few different techniques by
which the precision of the chatbot system can be made much
better. To make the responses given by the chatbot system
more meaningful and accurate the administrator has to train
the chatbot system with more information regarding to college
and increase the scope of knowledge base. Nevertheless,
gathering feedback from the potential user can be helpful in
developing the college Chatbot system, ultimately servicing
the user queries.

VI. REFERENCES
[1] A guide to Natural Langugage Processing, Available at
https://en.wikipedia.org/wiki/Natural_language_processing
[2] Chatbot definition, Available at https://medium.com/@mg/bot-is-a-
hilariously-over-simplified-buzzword-let-s-fix-that-
f1d63abb8ba7#:~:text=A%20chatterbot%20(also%20known%20as,via%
20auditory%20or%20textual%20methods.
[3] Introduction to Artificial Intelligence Markup Language, Available at
Fig. 6. Feedback System https://www.tutorialspoint.com/aiml/aiml_introduction.htm
chatbot response to the user queries. This feedback is stored in [4] Prof.K.Bala, Mukesh Kumar, SayaliHulawale, SahilPandita,“Chat-Bot
the database which can be used by college to know how For College Management System Using A.I” International Research
efficiently chatbot is answering user queries. Journal of Engineering and Technology (IRJET) Volume: 04, Issue: 11,
Admin need to provide username and password in order to Page no: 2030-2033| Nov 2017.
login. Only with proper authentication, admin is allowed to go [5] Porter Stemmer Algorithm, Available at
into the database. After admin provides proper username and http://snowball.tartarus.org/algorithms/porter/stemmer.html
password, admin gets logged into admin menu page as shown [6] Guruswami Hiremath, AishwaryaHajare, PriyankaBhosale,
RasikaNanaware, Dr. K. S. Wagh, “Chatbot for education system”
in Fig. 7. Where admin can perform operations such as add International Journal of Advance Research, Ideas and Innovations in
data to dataset, modify the existing data set, view all invalid Technology (IJARIIT) ISSN: 2454-132X, Volume: 4, Issue: 3, Page no:
queries, edit the predefined data, view user feedback, delete 37-43|2018.
the existing data and change password of admin module. All [7] Amey Tiwari, Rahul Talekar, Prof.S.M.Patil, “College Information Chat
the changes made here are directly changed in database. Bot System” International Journal of Engineering Research and General
Science (IJERGS) Volume: 5, Issue: 2, Page no: 131-137| March-April
2017.
[8] K. Jwala, G.N.V.G Sirisha, G.V. Padma Raju, “Developing a Chatbot
using Machine Learning” International Journal of Recent Technology
and Engineering (IJRTE) ISSN: 2277-3878, Volume: 8 Issue: 1S3, Page
no: 89-92| June 2019.
[9] Basics of Natural Language ToolKit, Avalilable at https://www.nltk.org/
[10] ] Naeun Lee, Kirak Kim, Taeseon Yoon, “Implementation of Robot
Journalism by Programming Custombot using Tokenization and Custom
Tagging” International Conference on Advanced Communications
Technology (ICACT) Page no: 566-570| Feb 2017.
[11] Fundamentals of Natural Language Processing - Tokenization,
Lemmatization, Stemming and Sentence Segmentation, Available at
https://colab.research.google.com/github/dair-
ai/notebooks/blob/master/_notebooks/2020-03-19-
nlp_basics_tokenization_segmentation.ipynb#scrollTo=H7gQFbUxOQt
b
[12] Jazzy spell checker Library, Available at http://jazzy.sourceforge.net/
[13] WordNet Algorithm, Avaliable at https://wordnet.princeton.edu/
[14] Setiaji Bayu, Wibowo Ferry “Chatbot Using a Knowledge in Database:
Fig. 7. Admin menu page Human-to-Machine Conversation Modeling” 7th International
Conference on Intelligent Systems, Modelling and Simulation (ISMS)
If the user is not satisfied with the chatbot responses then Page no: 72-77| Jan 2016. DOI: 10.1109/ISMS.2016.53.
he/she gives negative feedback. If admin finds the questions to [15] Synsets for a word in WordNet, Available at
be valid then admin can add answer to the particular query. If https://www.geeksforgeeks.org/nlp-synsets-for-a-word-in-
not then admin can delete the question, just by a single click. wordnet/#:~:text=WordNet%20is%20the%20lexical%20database,that%
20express%20the%20same%20concept.


 
     

Authorized licensed
View publication stats use limited to: Lovely Professional University - Phagwara. Downloaded on February 06,2021 at 10:07:14 UTC from IEEE Xplore. Restrictions apply.