About the projects
VulScrape is a vulnerability detection & prediction tool for forecasting exploits of common vulnerabilities found in source code written in C/C++. The tool is created as a Google Chrome extension for ease of use.
This project is inspired by the works of Li, Zhen, et al. (2021) & Fang, Yong, et al. (2020). The web extension integrates the vulnerability detection methodology from Li, Zhen, et al. (2018) where they used a deep neural network to detect code vulnerabilities and exploit prediction methodology from Fang, Yong, et al. (2020)'s ensemble machine learning algorithm.
All vulnerabilities that can be detected by VulScrape are listed under the National Vulnerability Database's CVE listing.
Partograph is a composite graphical record where measurements related to maternal and fetal data including uterine contraction, cervical dilatation, fetal heart rate, blood pressure, temperature, the descent of head, the color of membrane and liquor, molding of fetal skull bones, duration of labor, are plotted manually.
PartoCalc is an automated partograph generator mobile app that is operated by voice command. The app is intended for the use of nurses during the childbirth of pregnant mothers. This project was funded by Terre des hommes.
DSE stock estimator is a web application for predicting future stock prices of the Dhaka Stock Exchange using deep learning. This application can predict the opening and closing prices of the stock market companies currently present there. This application also features online model update for a stock market company and updates the closing and opening prices according to the latest entries of that company.
This project required Time-series forecasting and machine learning model integration with the web application. A temporary website was hosted during the demo of this solution and a complete user manual was done through Software Requirement Specification & Analysis.
Uncapacitated Facility Location Solver (UFLS)
A facility location problem is solved by finding the optimum placement of a facility in a given data set. A simple facility location problem is on general graphs is NP-hard to solve optimally, by reduction from the set cover problem The problem can be divided into 2 types of problems: Capacitated and Uncapacitated.
UFLS is a tool for solving uncapacitated facility location problems. 3 different approaches are available to solve the facility location problem:
Clustering Approach
Set Covering Approach
Iterative Approach