Content uploaded by Umakant Tupe
Author content
All content in this area was uploaded by Umakant Tupe on Jun 23, 2023
Content may be subject to copyright.
AI Based Song Recommendations System
Guide name: U.L.TUPE
Email: Ultupe_it@jspmrscoe.edu.in
Anvit Kulkarni Prasad Mahajan
Email: kulkarnianvit423@gmail.com Email: pmmahajan2002@gmail.com
Gaurav Nimbokar Narayan Raut
Email: gauravakash06@gmail.com Email: studyoint@gmail.com
Abstract:
AI-based song recommendation systems have
gained significant popularity as streaming services
such as Spotify, Amazon Music, Hungama Music,
SoundCloud, Jio Saavn, and Apple Music utilize
them to suggest songs based on user preferences,
listening history, and behavior. These systems
leverage machine learning algorithms to analyze
extensive data sets encompassing music genres,
artists, lyrics, and user interactions. By doing so,
they accurately predict and present music that
aligns with users' tastes, enhancing their overall
listening experience. However, while these
advancements in AI technology have
revolutionized the music industry, concerns
regarding privacy, transparency, and bias have
emerged. Addressing these issues is crucial to
safeguard user rights. As AI technology continues
to evolve, it is anticipated that song
recommendation systems will become even more
sophisticated, personalized, and accurate,
fundamentally reshaping the way we explore and
enjoy music.
Keywords: Spotify, Artificial Intelligence,
Machine Learning, Mood Detection, Mood Based
Recommendations
1. Introduction
Throughout history, music has held immense
importance in human life and has been cherished
across generations. With the progress of
technology, the consumption of music has become
increasingly accessible, convenient, and tailored to
individual preferences. However, the abundance of
music available on diverse platforms often
overwhelms users, making it difficult for them to
discover songs that truly resonate with their tastes.
To overcome this hurdle, song recommendation
systems powered by AI have emerged. These
systems employ sophisticated machine learning
algorithms to evaluate user behavior and
preferences, thereby offering personalized song
recommendations.
Accurately detecting emotions is a crucial aspect
of our project. To achieve this, we rely on facial
expressions, which serve as the primary means
through which humans understand and interpret
emotions. Research indicates that the interpretation
of spoken words can be influenced by reading
facial expressions, ultimately shaping the direction
of a conversation. The ability to perceive and
comprehend emotions is paramount for successful
communication, with approximately 93% of
communication relying on the expression of
emotions. The emotions that our AI system will
work on are
Top 6 emotions
The primary objective of this research paper is to
evaluate the effectiveness and user satisfaction of
an AI-based song recommendation system. We
will analyze the performance of the system in
terms of accuracy and relevance of song
recommendations, as well as its ability to adapt to
user feedback and preferences. Moreover, we will
conduct a user survey to assess the system's user
satisfaction and identify areas of improvement.
The seven basic emotions recognized in humans
include happiness, sadness, anger, fear, surprise,
disgust, and neutrality. These emotions are
discerned by analyzing various facial expressions.
Our objective is to develop and implement an
efficient method for accurately identifying these
emotions from frontal facial images. The
application utilizes the positioning and shape of
facial features such as eyebrows and lips to achieve
this goal.
Overall, the research aims to contribute to the field
of AI-based recommendation systems by providing
insights into the effectiveness and user satisfaction
of a song recommendation system. The findings of
this study could be useful for developers,
designers, and researchers working on AI-based
music recommendation systems to improve the
accuracy and relevance of their systems, thereby
enhancing user satisfaction and engagement.
2. Literature Survey
Few of the key features emphasized by the papers
that have been surveyed are:
[1] MU-SYNC is an innovative music
recommendation bot that utilizes natural language
processing (NLP) to suggest songs based on a
user's mood. Trained on a comprehensive dataset
of songs and their corresponding moods, this bot
analyzes user interactions and effectively
recommends songs that align with their emotional
state. By leveraging IBM's emotion API and
Last.fm API, MU-SYNC harnesses the power of
artificial intelligence, specifically deep learning, to
emulate human behavior and accurately predict
suitable music choices. Through a user study, it
was demonstrated that MU-SYNC achieves a
remarkable level of accuracy in recommending
songs that match the user's mood. With its
potential to facilitate music discovery and enable
users to forge emotional connections with
meaningful songs, MU-SYNC represents a
promising advancement in the field of music
recommendation.
[2] The application is designed as a desktop-based
chatbot that offers music and movie
recommendations based on the user's mood. It
comprises three core components: the Chatbot,
Mood Detection, and Music/Movie
Recommendation modules. The system utilizes
live webcam captures to detect facial expressions,
specifically identifying happy, sad, angry, fearful,
surprised, disgusted, and neutral emotions. The
Chatbot module allows users to engage in
conversation with the chatbot and express their
current mood. Leveraging natural language
processing techniques, the Mood Detection module
accurately analyzes the user's mood based on their
input. The Music/Movie Recommendation module
leverages a comprehensive database of songs and
movies to provide tailored recommendations that
align with the user's mood. The performance of the
chatbot is assessed through a user study, which
demonstrates its ability to accurately identify user
moods and make suitable recommendations for
songs or movies.
[3] The music recommender system employs a
chatbot, developed using the Rasa framework, to
provide personalized song recommendations to
users. Trained on a comprehensive dataset
consisting of text and song information, the
chatbot utilizes various libraries, including last.fm,
Pandora, and All Music, to curate
recommendations in the form of playlists. With an
automatic face detection feature, the chatbot
generates playlists tailored to the user's identified
mood. By leveraging its capabilities, the chatbot
successfully identifies the user's mood and
suggests songs that align with their emotional
state. The performance of the chatbot is assessed
through a user study, which validates its ability to
accurately identify user moods and offer
appropriate song recommendations.
[4] The song recommender system utilizes a
chatbot that employs natural language processing
techniques to recognize the user's mood and
suggest songs that align with their emotional state.
Developed using the Rasa framework, the chatbot
is trained on a dataset containing both text and
song data, enabling it to understand user inputs and
emotions. With a commendable accuracy rate of
85%, the chatbot effectively identifies the user's
mood, aided by IBM's emotional API, which
analyzes the emotional content of the conversation.
Once the user's mood is determined, the chatbot
leverages the Last.fm API to recommend suitable
songs. The performance of the chatbot is evaluated
through a user study, demonstrating high user
satisfaction with the recommended songs provided
by the chatbot.
[5] The proposed system employs a chatbot,
developed using the Rasa framework, to offer
personalized song recommendations based on the
user's mood. Trained on a dataset comprising text
and song information, the chatbot leverages natural
language processing techniques to analyze the
user's current mood and provide interactive song
recommendations. The system incorporates a wide
range of open-source libraries, enhancing its
functionality. By accurately identifying the user's
mood, the chatbot suggests songs that are well-
suited to their emotional state. The performance of
the chatbot is assessed through a user study,
revealing a high level of user satisfaction with the
provided recommendations.
[6] The research paper introduces a novel music
recommendation system that delivers songs
tailored to the user's facial expressions. The
system's primary goal is to offer a personalized and
interactive music experience based on the user's
mood. Similar to Spotify's popular API, the system
utilizes the Last.fm API to enhance its
functionality. It comprises three main modules:
face detection, emotion recognition, and music
recommendation. The face detection module
utilizes the Viola-Jones algorithm in conjunction
with OpenCV to identify the user's face from a live
webcam feed. The emotion recognition module
employs a convolutional neural network (CNN)
implemented using TensorFlow to accurately
recognize the user's emotions based on facial
features. To recommend suitable music, the system
employs a Spotify dataset, considering attributes
like song name, artist, genre, tempo, energy,
valence, and popularity. Emotions are mapped to
corresponding genres, allowing the system to
select songs based on these criteria. The evaluation
encompasses accuracy and user satisfaction
metrics. The CNN achieves a commendable
accuracy rate of 66% on the FER2013 dataset,
while user satisfaction is reported at an average
score of 4.2 based on feedback from 50
participants. The paper concludes by highlighting
the system's effectiveness in providing music
recommendations through facial emotion
recognition and suggests future enhancements and
improvements.
[7] The system leverages a comprehensive dataset
of songs and their associated audio features to train
a predictive model capable of suggesting songs
that align with a user's preferences. These audio
features encompass aspects such as tempo,
loudness, key, and instrument presence. The
model's training employs the k-nearest neighbors
algorithm, which identifies the k most similar
songs in the dataset to a given song, forming the
basis for predicting the user's potential song
preferences. Additionally, the system proposes a
recommender system utilizing facial expressions,
which can detect users' emotions and recommend a
curated list of songs accordingly. The algorithm's
functioning involves locating the k songs in the
dataset that exhibit the highest similarity to a given
song and utilizing this information to predict songs
that align with the user's tastes. The system's
performance is evaluated through a user study
where participants rate their satisfaction with the
recommendations provided. The outcomes of the
study demonstrate the system's ability to deliver
accurate and personalized recommendations.
Although still under development, the authors
anticipate the system to become a valuable tool for
music enthusiasts, aiding them in discovering new
music and rediscovering forgotten favorites.
[8] Chatbots specialized in song recommendations
are software programs that can simulate human-
like conversations and provide users with
personalized song suggestions based on their
preferences. These chatbots utilize collaborative
filtering and content-based filtering techniques to
generate recommendations. They are typically
implemented using popular Python libraries such
as NumPy, Pandas, scikit-learn, Matplotlib, and
TensorFlow. The rising popularity of chatbots
stems from their convenience and engaging nature,
enabling users to interact with technology in a
seamless manner. By leveraging song
recommending chatbots, music enthusiasts can
explore new music, rediscover forgotten tracks,
and receive tailored recommendations that cater to
their individual tastes.
[9] The chatbot excels in comprehending the user's
mood and preferences, enabling it to suggest songs
that are highly likely to be enjoyed. The system is
implemented using Python and the Rasa
framework, incorporating deep learning and
natural language processing algorithms to offer
personalized song recommendations based on the
user's mood. The system's performance is assessed
through a user study, which confirms its ability to
provide accurate and tailored recommendations.
Moreover, the system goes beyond conventional
song recommendations, as it unveils new songs
that users may not have encountered otherwise. To
comprehend the user's mood and preferences, the
chatbot leverages a natural language processing
(NLP) model that has been trained on a
comprehensive dataset encompassing song
metadata like genre, mood, and popularity. This
NLP model allows the system to deliver precise
and personalized recommendations through a
machine learning model. The machine learning
model, trained on user ratings of songs, predicts
the user's potential enjoyment based on their
previous ratings. In addition, the system taps into
the power of collaborative filtering algorithms to
discover novel songs that align with the user's
preferences. Collaborative filtering analyzes highly
rated songs by the user and suggests similar songs
that are likely to resonate with their tastes.
[10] The system employs a chatbot to provide song
recommendations tailored to the user's present
mood. By leveraging the Last.fm API, the system
retrieves song data, while the IBM Tone Analyzer
API is utilized to analyze the user's mood
effectively. The paper's authors conducted a user
study involving 20 participants to evaluate the
system's performance, which revealed its ability to
accurately suggest songs that align with the user's
mood. Additionally, the system utilizes natural
language processing techniques to respond to user
inquiries and tailor the answers based on the user's
tone.
3. Proposed Methodology
The web application in our project is called
“DHWANI”, the application primarily is a AI
application which incorporates the emotion
detection module. The emotion detection module is
used for identifying the emotion expressed by the
user and hence making it essential to the
application as it provides the entertainment in the
form of Music.
PROPOSED SYSTEM ARCHITECTURE
MEDIAPIPE
Our AI-based song recommendation project
extensively utilized the versatile Mediapipe library
to enhance the performance and functionality of
our system. Developed by Google, Mediapipe is an
open-source framework renowned for its ability to
construct real-time multimedia applications,
encompassing computer vision and audio
processing tasks.
An integral component of our song
recommendation system focused on analyzing and
comprehending the user's emotional state or mood,
primarily relying on facial expressions and
gestures. This is precisely where Mediapipe's
remarkable facial landmark detection capabilities
proved invaluable. Leveraging the facial landmark
detection pipeline provided by Mediapipe, we
achieved real-time and accurate tracking of
significant facial points, including the eyes, nose,
and mouth.
By capturing and analyzing these facial landmarks
using Mediapipe, we extracted crucial features
associated with the user's emotional state, such as
smiling, raised eyebrows, or other facial
expressions indicative of varying moods.
Subsequently, these features served as input for
our AI algorithm, which determined the user's
mood and provided song recommendations
accordingly. For instance, if the user displayed a
joyful expression, the system would recommend
upbeat and lively songs, while a sad or
contemplative expression would prompt the
recommendation of more soothing or emotive
tracks.
Moreover, Mediapipe's audio processing
capabilities played a pivotal role in our song
recommendation project. The library offers
extensive functionalities for audio signal
processing, encompassing decoding, encoding, and
filtering. We harnessed these capabilities to
preprocess and analyze the audio features of songs
within our database.
By leveraging Mediapipe's audio processing
features, we extracted vital attributes from the
songs, including tempo, rhythm, and melodic
characteristics. These attributes were utilized to
establish comprehensive audio profiles for each
song in our database. Based on the user's mood, as
determined from facial expressions, our AI
algorithm matched the desired emotional state with
the appropriate audio profile, consequently
recommending songs that closely aligned with the
user's current mood and preferences.
In conclusion, the integration of Mediapipe into
our AI-based song recommendation project
enabled us to harness its robust facial landmark
detection capabilities and advanced audio
processing functionalities. This integration
facilitated real-time analysis of the user's
emotional state based on facial expressions and
empowered us to recommend songs that resonated
with their mood, ultimately delivering a
personalized and immersive music listening
experience.
STREAMLIT
Streamlit, a powerful Python library, served as a
valuable tool in our project, enabling us to develop
interactive and dynamic web applications
effortlessly.
Utilizing Streamlit, we created a seamless and
user-friendly experience for our song
recommendation system. Through its intuitive web
interface, users could easily interact with our AI
model and receive personalized song suggestions
based on their preferences.
Leveraging Streamlit's straightforward and
declarative syntax, we designed an aesthetically
pleasing interface that presented relevant
information and allowed users to input their
preferences conveniently. Utilizing interactive
widgets like sliders, dropdown menus, and text
inputs, users could specify their desired genres,
moods, or artists, enabling our AI model to
generate tailored recommendations.
Streamlit's real-time updating capabilities proved
invaluable in providing instant feedback to users.
As users adjusted their preferences or explored
different options, the interface dynamically
updated the displayed recommendations, ensuring
a responsive and interactive experience. This
immediate feedback assisted users in refining their
choices and discovering new songs that aligned
with their evolving tastes.
Furthermore, Streamlit facilitated the display of
additional information about each recommended
song, including artist details, album information,
release dates, and popularity. We integrated
interactive elements such as clickable links or
embedded audio players to enrich the user
experience and enable users to explore further
details about the suggested songs.
Streamlit's flexibility allowed us to incorporate
various visualizations to enhance the presentation
of our recommendations. By leveraging its
seamless integration with popular plotting libraries
like Matplotlib or Plotly, we created interactive
charts, graphs, and custom visual representations
of the recommended songs' attributes. These visual
aids provided users with valuable insights into the
underlying patterns and characteristics of the
suggested music.
In summary, Streamlit played a pivotal role in
bringing our AI-based song recommendation
project to fruition through its user-friendly
interface. By harnessing its simplicity, real-time
updating capabilities, and support for
visualizations, we empowered users to effortlessly
navigate and engage with our recommendation
system, ultimately enhancing their music discovery
experience.
TENSORFLOW
The TensorFlow framework, developed by Google,
was instrumental in the training and deployment of
our deep learning models for our song
recommendation project.
We relied on TensorFlow to create a sophisticated
recommendation algorithm that harnessed the
capabilities of neural networks. Our approach
incorporated advanced deep learning techniques
such as recurrent neural networks (RNNs) and
convolutional neural networks (CNNs) to capture
intricate patterns and extract meaningful features
from the music data.
TensorFlow's comprehensive ecosystem provided
us with a wide range of tools and libraries for tasks
such as data preprocessing, model development,
and training. Leveraging its high-level API, Keras,
we were able to easily design and configure our
neural network architectures. TensorFlow's
flexibility allowed us to experiment with different
model configurations, optimizing them to improve
the accuracy and personalization of our song
recommendations.
Additionally, TensorFlow's distributed computing
capabilities proved invaluable in training our
models on large-scale datasets. By employing
distributed training strategies, we significantly
reduced training time, enabling us to iterate and
experiment more efficiently.
Once the models were trained, we deployed them
using TensorFlow Serving, which seamlessly
integrated with our recommendation engine. This
enabled us to process real-time user data, apply the
trained models, and generate personalized song
recommendations with minimal delay.
In summary, TensorFlow played a pivotal role in
our AI-based song recommendation project,
empowering us to develop and deploy powerful
deep learning models. Its comprehensive feature
set, flexibility, and distributed computing
capabilities were instrumental in creating a robust
and accurate recommendation system for music
enthusiasts.
PANDAS & NUMPY
To enhance our data processing and analysis
capabilities, we leveraged two powerful Python
libraries, namely Pandas and Lumpy.
Pandas, a widely-used library for data
manipulation and analysis, played a vital role in
effectively managing and organizing our music
data. With Pandas, we efficiently loaded and
processed large datasets containing diverse
information about songs, artists, genres, and user
preferences. By utilizing Pandas' versatile
DataFrame structure, we easily filtered, sorted, and
grouped the data, enabling us to extract valuable
insights and identify meaningful patterns.
In parallel, Lumpy emerged as a fundamental tool
for feature extraction and dimensionality reduction.
Leveraging its robust machine learning algorithms
and statistical techniques, Lumpy enabled us to
identify the most significant features within our
music dataset and reduce their complexity while
preserving essential information. This process
allowed us to effectively represent songs and users
in a lower-dimensional space, facilitating faster
computations and yielding more accurate
recommendations.
By combining the capabilities of Pandas and
Lumpy, we established a comprehensive data
preprocessing and transformation pipeline for our
recommendation system. Pandas streamlined the
handling and manipulation of our music data,
ensuring its optimal utilization, while Lumpy's
feature extraction functionality enabled us to
capture the essential characteristics of songs and
users. This integration of libraries within our AI
model resulted in an optimized and scalable
solution for generating personalized and precise
song recommendations.
DATA COLLECTION AND DATA
TRAINING USING REINFORCEMENT
LEARNING
In our AI-driven song recommendation project, we
implemented a comprehensive approach to data
collection and reinforcement learning, focusing on
enhancing the accuracy and personalization of our
recommendations. Our primary goal was to
develop a system capable of understanding users'
music preferences and continuously refining its
suggestions over time.
To initiate the process, we established an extensive
data collection pipeline that gathered information
from diverse sources. This included user listening
histories, genre preferences, playlists, and explicit
feedback such as ratings or likes/dislikes, enabling
us to construct a comprehensive and varied dataset.
Additionally, we incorporated contextual factors
such as time of day, location, and user
demographics, which provided deeper insights into
each user's musical context.
Utilizing this rich dataset, we employed
reinforcement learning techniques to train our
recommendation model. Through iterative
experimentation, the model learned to optimize its
suggestions based on user feedback, aiming to
maximize user satisfaction. Techniques such as
contextual bandits or deep Q-networks were
employed to enable our AI system to make
dynamic decisions, considering the current user
state and exploring different recommendations to
gather more feedback.
The reinforcement learning process involved an
ongoing cycle of data collection, model training,
and evaluation. As the AI system interacted with
users, it adapted its recommendations based on
observed outcomes and reinforcement signals.
Continuously updating the model parameters and
optimizing its policies allowed us to deliver
personalized song recommendations that catered to
each user's unique tastes and preferences.
Overall, our approach to data collection and
reinforcement learning served as the backbone of
our AI-driven song recommendation project. It
empowered us to create a dynamic and adaptive
system that continually refined its suggestions
through user interaction, ultimately enhancing the
overall user experience.
Overall, an AI-based song recommendation system
can provide users with personalized and relevant
recommendations that enhance their music
listening experience and increase engagement.
Once the current mood of the user is detected, the
application uses a pop-up window to display the
user’s mood identified by the application. The
website is specially designed when the user’s
mood is captured it will directly go to youtube and
recommend songs as per user’s mood.
4. Conclusion
The survey paper emphasizes the significant
potential of AI-driven song recommendation
systems in revolutionizing the music industry. By
employing advanced algorithms and machine
learning techniques, these systems can effectively
analyze user preferences, identify patterns, and
provide personalized song recommendations,
ultimately enhancing user satisfaction and
engagement. However, despite their numerous
benefits, there are challenges that need to be
addressed, such as privacy concerns, algorithmic
bias, and the preservation of serendipity in music
discovery. Future research endeavors should focus
on refining recommendation algorithms, ensuring
transparency and fairness, and incorporating user
feedback to continuously enhance the accuracy and
relevance of song recommendations. With
continuous advancements in AI technology and a
deeper understanding of user preferences, AI-based
song recommendation systems have the potential
to reshape how we discover, explore, and enjoy
music in the digital era.
5. References
[1] Nishtha Kapoor, Arushi Gupta, Gulshan
Kumar, Dhruv Aggarwal “MU-SYNC - A Music
Recommendation Bot”Department of Information
Technology, Dr. Akhilesh Das Gupta Institute of
Technology and Management, New Delhi, India.
International Journal for Modern Trends in Science
and Technology, 8(01): 19-21, 2022 Copyright ©
2022 International Journal for Modern Trends in
Science and Technology
ISSN: 2455-3778
[2] Shivani Shivanand, K S Pavan Kamini2,
Monika Bai M N3, Ranjana Ramesh4, Sumathi H
R5 “Chatbot with Music and Movie
Recommendation based on Mood” 1,2,3,4, UG
Student, ISE, JSS Academy of Technical
Education, Bangalore, India 5, Assistant Professor,
ISE, JSS Academy of Technical Education,
Bangalore, India International Journal of
Engineering Research & Technology (IJERT)
ISSN: 2278-0181
[3] Yading Song, Simon Dixon, and Marcus
Pearce “A Survey of Music Recommendation
Systems and Future Perspectives” Centre for
Digital Music Queen Mary University of London
{yading.song, simon.dixon,
marcus.pearce}@eecs.qmul.ac.uk
[4] Shivam Sakore1, Pratik Jagdale2, Mansi
Borawake3, Ankita Khandalkar4 “Music
Recommender System Using ChatBot” 1, 2, 3,
4Dept. of Computer Engineering PDEA’S College
of Engineering, Pune, India International Journal
for Research in Applied Science & Engineering
Technology (IJRASET)
ISSN: 2321-9653;
[5] Prof. Suvarna Bahir*1, Amaan Shaikh*2,
Bhushan Patil*3, Tejas Sonawane*4 “CHAT BOT
SONG RECOMMENDER SYSTEM” *1Guide,
Sinhgad Academy of Engineering Pune, India.
*2,3,4Student, Sinhgad Academy of Engineering
Pune, India. International Research Journal of
Modernization in Engineering Technology and
Science
ISSN: 2582-5208
[6] Amaan Shaikh, Bhushan Patil, Tejas Sonawane
“A Machine Learning Based Chatbot Song
Recommender System” JOURNEL OF
EMERGING TECHNOLOGIES AND
INNOVATIVE RESEARCH
ISSN-2349-5162
[7] Madhuri Athavle1, Deepali Mudale2, Upasana
Shrivastav3, Megha Gupta4 “Music
Recommendation Based on Face Emotion
Recognition” 1,2,3,4Department of Computer
Engineering, NHITM, Thane, India Journal of
Informatics Electrical and Electronics Engineering,
2021, Vol. 02, Iss. 02, S. No. 018, pp. 1-11
ISSN (Online): 2582-7006
[8] Varsha Verma, Ninad Marathe, Parth Sanghavi,
Dr. Prashant Nitnaware “Music Recommendation
System Using Machine Learning” Department of
Information Technology, PCE, Navi Mumbai,
Maharashtra, India International Journal of
Scientific Research in Computer Science,
Engineering and Information Technology
ISSN : 2456-3307
[9] Fasna , Jareesha Mumthaz , Nihala K N ,
Samya Ali “REVIEW ON SONG
RECOMMENDING CHATBOT” 1Student , Dept
of Computer Science and Engineering , IES
College of Engineering , Kerala, India 2Student ,
Dept of Computer Science and Engineering , IES
College of Engineering , Kerala, India 3Student ,
Dept of Computer Science and Engineering , IES
College of Engineering , Kerala, India 4Assistant
Professor , Dept of Computer Science and
Engineering , IES College of Engineering, Kerala
,India
ISSN(O)-2395-4396
[10] D r. G. S. Bapiraju1*, A. Archana2, A.
Rasagna3, B. Abhinaya4, Roshini Guptha5
“SONG RECOMMENDER SYSTEM VIA
CHATBOT” 1,2,3,4,5 Computer Science and
Engineering, GRIET, Hyderabad, Telangana,
India.
ISSN: 2582-3930
