Thesis Title

Utilizing Machine learning and Deep learning to Analyze Public

Perceptions of Bangladeshi Cricketers: An In-depth Examination of
Facebook Bangla Comments

Supervisor:
Team Members:
Mahbuba Yesmin Turaba
1. Sharika Jahan Aunto (03/20/50/3201)
Lecturer
2. Nusrat Jahan Smriti (03/20/50/3094)
Department of CSE,
3. Masum Biswas (03/20/51/3273)
World University of Bangladesh
ame_surname (STUDENT ID)

Master’s Thesis_or title in short

insert date 1/9
 Outline

● Introduction
● Objectives
● Justification of Study
● Scope of Study
● Dataset sample
● Literature Review
● Conclusion
● References

January 16, 2024

2 / 13
Introduction

Cricket is one of the most popular game and Bangladeshi cricket team is well-recognized in
worldwide. It is more than just a game, holds a special place in the hearts of Bangladeshis.
Therefore, it is expected that there are passionate fan-base, player adoration, social media
engagement and cultural impact.

In our paper, the strategy is to extract real time thought

of public opinion about Bangladeshi cricketers using
Facebook Bangla comments through sentimental
analysis.

January 16, 2024

3 / 13
 What is Sentimental Analysis?

● Sentimental analysis is a standard method to determine human emotion.

● Natural Language Processing (NLP) is primarily utilized for sentiment analysis which is a
mining method that determines a text's various feelings.
● These emotions could be classified as various categories such as happy, sad, angry and neutral.

January 16, 2024

4 / 13
 Objectives

● To prepare a Bangla text-based dataset from cricketers verified

Facebook profile.
● To analyze sentiment in Bangla text using machine and deep
learning algorithms, based on comments and descriptive analysis.
● To gain a deeper understanding of people's view and capture real-
time scenarios regarding Bangladeshi cricketers, as well as to
predict three emotions from Bangla text.

January 16, 2024

5 / 13
 Justification of Study

In recent times, around 46 million people, actively engage with the social media platform such as Facebook, Instagram,
Twitter and give their insights through these platforms [2]. Here, sentimental analysis is justified for-

1. Insightful Sports Analytics: Analyzing public perceptions through social media comments can offer unique insights
into how fans view and react to Bangladeshi cricketers. This information can be utilized by cricket analysts and teams
to enhance their understanding of player popularity, strengths, and weaknesses.
2. Data-Driven Decision Making: The use of machine learning and deep learning techniques enables the extraction of
actionable insights from the large volume of Facebook Bangla comments. This information can inform decision-
making processes for cricket boards, sponsors, and players, leading to more strategic and informed decisions.
3. Player Development: Understanding how fans perceive cricketers can be valuable for player development programs.
Identifying areas of improvement or aspects that resonate positively with the audience can guide training strategies and
career development for players.
4. Social Media Strategy: The findings from our analysis can be used to tailor social media strategies for individual
players or the cricket team. Knowing what type of content generates positive engagement and resonance among fans
can enhance the team's online presence.

January 16, 2024

6 / 13
 Scope of Study

The working procedure of our paper has been summarized below-

● We developed more than 5000 comments for our dataset by both scrapping and manually from six
Bangladeshi cricketers profile.

● We classified our 5000 comments into three categories such as positive, negative and neutral.

● We conducted a descriptive analysis of demographic data to examine the characteristics of individuals

engaged in spreading of real time opinion, hate speech and cyberbullying.

● We will accomplish experiments with our dataset by applying deep learning and machine learning
algorithms for the purpose of categorizing various sentiments.

January 16, 2024

7/ 13
 Dataset Sample

Comments Sentiments Like Haha Angry Love Care Sad Wow

এশিয়া কাপ কিংবা বিশ্বকাপ যেকোনো একটি ট্রফি সাকিব আল হাসানের নেতৃ ত্বে আমরা Positive 0 0 0 0 0 0 0
দেখতে চাই তবেই ফ্যান ফলোয়ারদের মনের আশা পূর্ণ হবে

ফেসবুকে আজ অনেকে আবেগময় কথা বলতেছে কিন্তু এদের মধ্যে ৯৫% মানুষ Neutral 5 0 0 0 0 0 0
তামিমকে নিয়ে ট্রোল করছে

একটা জাতীয় দলের ক্যাপ্টেন এর কমনসেন্স কোন জায়গায় পৌছালে এইরকম মার্কে টিং Negative 1 0 0 0 0 0 0
করতে পারে তা ভেবে দেখা উচিতবয়স বাড়ার সাথে সাথে তার পেশাদারিত্ব তলানিতে
গিয়ে ঠেকেছে এবারো মনে হয় বিশ্বকাপে ডাব্বা মারার পরে বললে লাঞ্চ ভালো ছিল তাই
আউট হয়ে গেছি

January 16, 2024

8 / 13
 Literature Review
S.no Author & Year Paper Name Used Algorithms Result / Best Remarks
Accuracy
[1] M. S. S., Rafa, S. R., FinTech: Deep Learning- BiLSTM, CNN, LSTM BiLSTM- 97.07% Larger dataset containing English and
Abir, A. E. H., and Based Sentiment Bangla combining reviews are recommended
Das, A. K. (2023) Classification of User for better result.
Reviews from Various
Bangladeshi Mobile
Financial Services

[2] Khan, M. S. S., Rafa, Sentiment Analysis on Used different machine learning SVM- 62% Dataset can be more focused in terms
S. R., Abir, A. E. H. Bengali Facebook techniques to train the dataset. RF- 58% of prediction. More machine learning
KNN- 55%
and Das, A. K. (2021) Comments To Predict techniques can be applied in future.
Fan’s Emotions Towards a
Celebrity

[3] A. N. Tusher, S. Islam, User Perspective Bangla Six state-of-the-art machine learning MNB- 73.27% Data collection phase and lack of legitimate
M. T. Islam, M. S. R. Sentiment Analysis for classification algorithms viz. Decision opinions are the main challenges.
Sammy, M. S. Rahman Online Gaming Addiction Tree, Random Forest, Multinomial
and M. S. Sadik using Machine Learning Naive Bayes, Extreme Gradient
(2022) Boosting, Support Vector Machine
and K Nearest Neighbor are used to
train the model.

January 16, 2024

9 / 13
 Literature Review
S.no Author & Year Paper Name Used Algorithms Result / Best Remarks
Accuracy
[4] Sarkar, O., Ahamed, An Experimental CNN, LSTM, BiLSTM BiLSTM- 83.77% Performance would increase if the model was
M. F., Khan, T. T., Framework of Bangla Text trained by more information.
Ghosh, M. K., and Classification for
Islam, M. R. Analyzing Sentiment
Applying CNN &
BiLSTM

[5] Younis, Eman. (2015) Sentiment Analysis and Clustering,Classification, The mean Sentiment For further improvement, un-supervised
Text Mining for Social associations, visualizing score for Tesco is techniques could be used in this opinion
Media Microblogs using 0.1595. mining.
Open Source Tools: An
Empirical Study.
International Journal of
Computer Applications

January 16, 2024

10 / 13
 Conclusion

❏ Our study employed a dataset including over 5000 comments extracted from the Facebook posts of
six professional cricketers.
❏ This dataset will be examined by machine & deep learning methodologies include XGboost, LSTM,
LR, CNN, Linear SVM algorithms to effectively categorize comments into several categories,
including positive, negative, or neutral.

January 16, 2024

11 / 13
 References
[1] Al Ryan, A., Mahmud, M.S., Mahi, H.H.C., Hossen, M.S., Shimul, N.I. and Noori, S.R.H., 2023, February. FinTech: Deep
Learning-Based Sentiment Classification of User Reviews from Various Bangladeshi Mobile Financial Services. In
International Conference on Computational Intelligence in Data Science (pp. 126-140). Cham: Springer Nature Switzerland.

[2] Khan, M.S.S., Rafa, S.R. and Das, A.K., 2021. Sentiment analysis on bengali facebook comments to predict fan's emotions
towards a celebrity. Journal of Engineering Advancements, 2(03), pp.118-124.

[3] Tusher, A.N., Islam, S., Islam, M.T., Sammy, S.R., Rahman, M.S. and Sadik, M.S., 2022, November. User Perspective
Bangla Sentiment Analysis for Online Gaming Addiction using Machine Learning. In 2022 Sixth International Conference on I-
SMAC (IoT in Social, Mobile, Analytics and Cloud)(I-SMAC) (pp. 538-543). IEEE.

[4] Sarkar, O., Ahamed, M.F., Khan, T.T., Ghosh, M.K. and Islam, M.R., 2021, May. An experimental framework of bangla text
classification for analyzing sentiment applying CNN & BiLSTM. In 2021 2nd International Conference for Emerging
Technology (INCET) (pp. 1-6). IEEE.
[5] Younis, E.M., 2015. Sentiment analysis and text mining for social media microblogs using open source tools: an empirical
study. International Journal of Computer Applications, 112(5).

January 16, 2024

12 / 13
Thank you