ACADEMIC RESEARCH · B.SC. LEVEL 4

Sinhalese Language Based Hate Speech Detection and Intent Analysis on Tweets

Research into detecting Sinhalese-language hate speech and analysing tweet intent using semantic analysis, text preprocessing and data-mining techniques.

Python NLTK Pandas NumPy Scikit-learn PyCharm
View Project Slides ↓ ← Back to portfolio

01 / RESEARCH OVERVIEW

Turning Sinhala text into meaningful insights.

The research explored the development of a tool for detecting Sinhalese-language hate speech and analysing the intention of tweets.

Research focus

Semantic analysis and data-mining techniques were used to address the challenge of understanding and classifying Sinhala-language social media content.

My contribution

My contribution included implementing a text preprocessing module whose outputs were used as inputs for analysis. I also worked on classifying Sinhala texts found on Twitter into primary domains and subjective categories.

Research objective

The classification process aimed to support further analysis of hate according to domain and subject classes, helping define the severity level of tweet content.

02 / RESEARCH MODULES

Breaking down the classification workflow.

The original project describes multiple stages for processing and classifying Sinhala Twitter text.

MODULE 01

Twitter text analysis

Working with Twitter text as the source material for the research.

MODULE 02

Text preprocessing and classification

Preparing Sinhala text and categorising tweet content by domain and subject.

MODULE 03

Hate detection and intent analysis

Applying classification outputs to further analysis of hate and the intent of tweet content.

DETAILED CONTRIBUTION

Text Classification Module

Researched Sinhalese Language Based Short Text Classification to Improve Information Filtering. The module covered extracting Sinhala Twitter texts, preprocessing them, and categorising tweet content according to subject and domain.

Original project material · 01
Original project material · 02