Scribd
Upload
2 views1 page

Problem Statement Taha

The document discusses the increasing issue of cyberbullying in the Bangla language, highlighting the need for improved detection methods using NLP techniques. It presents a study that achieved an accuracy of 87.91% with binary classifications and 79% for harassment detection using multiclass classifications. The paper emphasizes the necessity of incorporating image analysis and addresses limitations such as the need for a comprehensive Bangla dictionary and challenges in understanding complex sentences.

Uploaded by

mtaha17.cse
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
2 views1 page

Problem Statement Taha

The document discusses the increasing issue of cyberbullying in the Bangla language, highlighting the need for improved detection methods using NLP techniques. It presents a study that achieved an accuracy of 87.91% with binary classifications and 79% for harassment detection using multiclass classifications. The paper emphasizes the necessity of incorporating image analysis and addresses limitations such as the need for a comprehensive Bangla dictionary and challenges in understanding complex sentences.

Uploaded by

mtaha17.cse
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

Cyber bullying for the Bangla language with Image macro: an NLP based

approachment.

Problem statement:
As the time is going on, the English language is developing itself day by day. With this issue and the wide
using of social networks, the the rate of the malpracticing of those is getting higher. So, it’s the time to
review the classical system of indicating cyber bullying system and may also introduce some new ways,
which has the possibility of work different way.

Problem: detecting several types of cyberbullying from the web based texts.

What are the assumptions:

Previously researchers used various types of technology such as CNN, SVN separately or combindly and
try of estimate the heights accuracy of the both results. Also used RNN method. Here in this paper, they
used Binary and Multiclass classifications.

Contributions:

Finally almost 87.91% with the binary classifications, 79% for the harassment with the multiclass
classification of accuracy and so on.

Why is it still problem:

Its mandatory to add the picture analysis.

Limitations:

As far as the challenge, is to make a large accurate Bangla dictionary and some other problems just like
to identify the complex sentences.

You might also like