// Data Science Project

SiasatAI
Misinformation Detection

Bilingual text classification application optimized to identify local rumors and news reports in Bahasa Malaysia and English.

01.The Problem

The rapid spread of localized misinformation and rumors in both Bahasa Malaysia and English required an automated, highly-accurate detection system capable of sub-2ms response times.

02.The Solution

Engineered an end-to-end pipeline. Automated text data extraction via Python web scraping to build a clean dataset of 64,000+ records. Selected and trained a regularized Logistic Regression model achieving 93.99% validation accuracy, and deployed it via FastAPI on a Vercel Serverless CDN.

System Data

RoleFull-Stack AI Developer

StatusLive Active

Tech Stack

PythonFastAPIScikit-learnVercelWeb ScrapingNLP

View Project Source Code

SiasatAI Misinformation Detection

01.The Problem

02.The Solution

System Data

SiasatAI
Misinformation Detection