NEURAL NETWORK BOT

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 07 | July 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 495
NEURAL NETWORK BOT
Praneeth Finney .B1, Sai Ganesh Angadi2, Aparna .K.S3
1,2 B.E student,Dept Of Computer Science, Rao Bahadur.Y.Mahabaleswarappa college, Ballari,Karnataka,India
3Assistant Professor, Department Of Computer Science, Rao Bahadur.Y.Mahabaleswarappa college
Ballari,Karnataka,India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - The most commonly known assistant which
digitally interacts is the Chat bot which uses a technology for
interaction between the human and the computers, this
project has a bot based on neural networking and trained for
analysing the queries of the users and simultaneously
analysing the emotions of the user since emotions play a
crucial part in analysis and to provide proper response. These
maybe in the form of speech, gesture or signs , Moreover with
the increased usage of messenger or texting applications,
emotions from the text is extracted. Specially targeting
youngsters who are feeling low. these Bots play the role of
good companion ,creating a strong relationship between the
users and the bot and even moving a step further provide
suggestions and help the users
Key Words: Chatbot, emotion recognition,neural network,
Natural Language processing.
1. INTRODUCTION
In simple terms Chatbot can be termed as
software/programs that can interact, communicate with
humanbeingsandrespondusingnaturallanguageprocessing
and neural network. As a conversational partner, algorithms
convincingly simulate humans beings .These simulated
programs so called chatbots are classifies into two fairly
based on their response mechanism and training. The First
category of chatbot comes under those botsthatrespondand
process the queries based on the a fore mentioned set of
training examples and rules, hence these are bound to some
extent hence called as retrieval based.
Second type of bot is the one dynamic in nature and
adapts and learns as time progresses, so called as generative
bots. These are further classified as:
 Open domain with retrieval-based responses
 Open domain with generative-based responses
 Closed domain with retrieval-based responses
 Closed domain with generative-based responses
Machine and human interaction has been done through a
process called the natural language processing and few
example of bots that are contemporary which use such
processing are Siri, cortona, allo etc.
Speaking of emotions and life ,emotions are an integral
part of life or can be termed as life itself in some cases since
they are found to have an impact on various aspects namely
thoughts,inspiration,,imagination,focus,,experienceswhich
cumulatively influence the behaviour and decisions of a
person.
Our project and approach is focused on casual
interactions with users usingthebot,usersoftenshowawide
range of emotions, such as being nervous about studying,
being desperate for a new career, feeling depressed about
wanting to break up, etc. Understanding these emotions and
developing an emotionally attentive approach enable a deep
sustaining relationship with users and to know their
emotional desires. Many students are turning to the Internet
to release their negative feelings, and it is shown to have
positive effects on loneliness, depression, and stress.
Numerous youngsters especially students and fresh job
seekers are going to the Internet to deliver their gloomy
sentiments, and it is displayed to emphatically affect
dejection, sadness, and stress.
In reality this bot can be called as a virtual friend to them
who want a friend in deed but can’texpressopenheartedlyto
the real world but want to express irrespective of what can
have this bot as a companion, wherein the bot poses few
questions and interacts to understand the emotions and
problem of the clients/users purely based on the chat
interaction between the bot and the user using the following
modules specified in
2. Modules Description
A. Random:
This is a Built-in module of python and can be used to
generate random numbers or perform actions such as
printing a random value in case of a stringorstructuredonor
unstructured structured list.Basically it generates or creates
random numbers.
B. Json
Primarily Json is a Built-in function to work with Json data in
python. Json is an open standard file format at the same time
it is an interchangeable format too. It consists of data objects
carrying the pair and even the data type.

C. Pickle
This Pickle module deals with conversion of python objects
structures into binary form, that is it converts objectssuchas
list dictionary etc. Into 0s and 1s which is known as
serializing and de serializing. Objects are converted into a
stream of bytes known as character stream.
D. Sys
This module is about the runtime environment of python,
providinginformationanddetailingaboutmethods,functions
and various constants related to the interpreter which
manipulates the environment and maintains the functions
interacting with them.
E. Numpy
Numpy abbreviated as numerical python deals with various
arrays multidimensional in nature and the collection for
processing thesame.Differentdomainssuchaslinearalgebra
and matrices can be performed; even logical operations and
mathematical operations can be performed on the arrays.
This enables efficiency when dealing the logic and
mathematics in terms of arrays especially.
F. NLTK
NLTK module is a toolkit for natural language processing
that is it makes natural language understandable to
machines. Generally the inputs or the text given is in high
level language which is not machine understandable such
unstructured data isconvertedusingtheNLTKknownaspre-
processing . This pre-processing becomes much easier with
NLTK as it has a large number of libraries and even provides
data sets which can be used for testing.
G. Tensorflow
Tensorflow is an accumulation of tools, libraries and
resources for the usage ofcommunity which is dynamic with
different ecosystems .This module was created for machine
learning processes and is open sourced for model training
and easy debugging which makes execution easier and
intuitive. Tensor flow is not language restricted and can be
deployedand adapted to any system making productionalot
easier.
It helps in the creation of deep learning models for efficient
and fast paced computing ability especially in case of
numerical.
H. Speech recognition
Speech recognition in pythonisamodulewhichidentifiesthe
spoken words in and modifies them into text making it
machineunderstandable andprocessable. In simple termsit
allows the software to acquire input from the Mic and
analyses it, processes it and provides a suitable response.
Moreover speech recognition supports several search
engines and API
I. Pywhatkit
Pywhatkit library is a type of automation module, especially
for automation of messages in WhatsApp. It simplifies
sending messages using a minimal code written in python.
Pywhatkit is popular because of its user friendly set up and
easy to use library functions and features.
Not just messaged but even other media such as playing
YouTube videos, sending mails, sending imaged and files can
be done through this.
J. LRU CACHE
The functools module has the LRU cache, this tools uses
memorization as a technique in order to reduce the time of
execution to a minimum, making the application faster in
processing. LRU cache keeps track of all the bits and
maintains the least used bits or data, known as optimization
of data helping to the full computational ability of the
program or computer.
K. Pyttsx3
This Is a library function of python ,this library function is
used for conversion of text to speech that is it converts the
data typed into a voice audible as human.
The program converts into audible voice either male or
female voice, a unique feature of this library is that it works
even when there is no internet connectivity i.e offline.
L. Selenium
This Selenium is a test driver in python used for automation
which is the scripts related to browsers are automated. The
interaction with web browsers such as chrome, Firefox etc.
are made easy and efficient through API.
It is a testing tool wherein it converts the test cases provided
into scripts for automation which is quite flexible and
supports multiple operating systems and languages.
M. Lemmatizer
Lemmatizer is a module in the NLTK library,usedfornatural
language processing. This process of lemmatization involves
grouping of similar kinds of words into one item for better
understanding since they group the words having relatively
similar meaning and structure.
N. Pyaudio
Pyaudio is a module that can be used across different
platforms for input/output library. This module helps in
recording and playing different types of audio in platform.

O. Matplotlib
Matplotlib is a multi-platform library used for visualization
of data into arrays for 2d plotting. It creates plots which
maybe static or animated making easier data visualization
and analysis.
P. Seaborn
This Seaborn is a python library used for high level of data
visualization and is used on top of Matplotlib. Exploratory
data analysis in terms of graphs is done through this by
drawing graphs that are informative for data analysis.
Randomly distributed data isanalysed andgraphsaredrawn
through Seaborn.
3. INPUT AND OUTPUT
Input is taken from the user in the form of speech, facial
recognition that is in the form of picture or in the video
format and analysed. Based on this input received the bot
tries to analyse the scenario and condition and emotional
state of the client or the user and gives output based on the
same as suggestions such as encouraging to read a book
suggested ,music recommendation, movie recommendation
or some suggestive measures.
4. ARCHITECTURE
The architecture shown in figures 1 and 2 represent the
various logical steps in the natural language processing and
the steps involved in the training ofdatasetandusageof Json
file for the same which forms the crux of the neural bot
implementation.
The Feed forward model is adopted and the data that is
encoded is used in the training stage, a number of input
layers, hidden layers and output layers are created in
accordance to trainingdata formingthe neural network.This
is called the Natural language Pre-processing whichincludes
Tokenization, Stemming/Lemmatization, and Regex
Extraction.
After the pre-processing stage the interface moves to the
prediction phase where the different queries of the
user/client are taken. Querying is done using UI and
forwarded to the query pre-processing stage wherein the
queries are encoded and passed on to the processing and
prediction stage. In this stage the queries are analysed
decoded and gives back the suitable or appropriate
suggestions based on the query to the user.
5. CONCLUSION AND OUTPUT
This paper has shown how neural networking using
tensor flow and other modulesinpythonsuitablycanbeused
to build a neural network bot that takes input from the user
in the form of text such as messaging andinteractionwiththe
bot, speech and video or picture. The main agenda of this bot
was to take the inputs from the people especially from the
younger peers such as students or the young working class
prone to different traumatic situations and those who do not
want to share or can’t share their feelings with ease, this bot
serves as a virtual companion and helpsbytakingthevarious
inputs, analysing the mood and provides the necessary
feedback.

This bot can be further enhanced by integrating the social
media accounts of the user taking their consent and
dynamically analyse the mood of the person at regular
intervals and provide futuristic suggestions based on the
mood of the person to cheer up the user such as movie,
music, book, restaurant or even provide suggestions to talk
to necessary people in utmost necessity. In future in this
growing social media having been showing the signs of least
interaction among people and this needs to be taken care.
Hence this bot can be upgraded to bring out the human
nature and basic human traits such as being social,
interacting etc. Not just restricting it to an application this
can be integrated and made cross platform application
REFERENCES
[1] P. Kumar, M. Sharma, S. Rawat and T. Choudhury,
"Designing and Developing a Chatbot Using Machine
Learning," 2018 International Conference on System
Modeling & Advancement in Research Trends (SMART),
2018, pp. 87-91,doi:10.1109/SYSMART.2018.8746972.
[2] M. Karna, D. S. Juliet and R. C. Joy, "Deep learning based
Text Emotion Recognition for Chatbot applications,"
2020 4th International Conference on Trends in
Electronics and Informatics (ICOEI)(48184), 2020, pp.
988-993, doi: 10.1109/ICOEI48184.2020.9142879.
[3] F. Patel, R. Thakore, I. Nandwani and S. K. Bharti,
"Combating Depression in Students using an Intelligent
ChatBot: A Cognitive Behavioral Therapy," 2019 IEEE
16th India Council International Conference (INDICON),
2019, pp. 1-4, doi:
10.1109/INDICON47234.2019.9030346.
[4] P. Kumar, M. Sharma, S. Rawat and T. Choudhury,
"Designing and Developing a Chatbot Using Machine
Learning," 2018 International Conference on System
Modeling & Advancement in Research Trends (SMART),
2018, pp. 87-91,doi:10.1109/SYSMART.2018.8746972.
[5] M. T. Mutiwokuziva, M. W. Chanda, P. Kadebu, A.
Mukwazvure and T. T. Gotora, "A neural- network based
chat bot" 2017 2nd International Conference on
Communication and Electronics Systems (ICCES),
Coimbatore, pp.211– 217,2017
[6] G. K. Vamsi, A. Rasool and G. Hajela, "Chatbot: A Deep
Neural Network Based Human to MachineConversation
Model," 2020 11th International Conference on
Computing, Communication and Networking
Technologies (ICCCNT), 2020, pp. 1-7, doi:
10.1109/ICCCNT49239.2020.9225395.
[7] Akshay Kumar, Pankaj Kumar Meena, Debiprasanna
Panda, Ms. Sangeetha, “Chatbot In Python”, SRMIST,
Chennai

NEURAL NETWORK BOT

Recommended

More Related Content

Similar to NEURAL NETWORK BOT (20)

More from IRJET Journal (20)

Recently uploaded (20)

NEURAL NETWORK BOT