Revamped Malicious URL Detector, that increased its accuracy by 25% and has enabled Continuous Delivery and a Toolchain which can be used to automate builds, tests, deployments. Django REST framework is used here that uses a REST API call.
Concept Video of Application
3. A.I. Powered Safe Browsing Extension (Chromium and Firefox based)
ACTIVE USERS IN 22 COUNTRIES INCLUDING (US, FR, IN, GB)
Source: addon's statistics
About this Project
● Always-on detection system (Browser Extension) to provide protection within limited internet connectivity restricted networks, or while connected to a risky WIFI network.
● Option to share generated Blacklist Database to organisation’s administration to block sites from blacklisted registrars, emails, domains for enhanced security.
● Web Browser extension that consumes very little data and is extremely fast.
● Machine Learning backend based on dynamic features like WHOIS, ALEXA RANK to predict safety status of a URL.
● Django REST API that is linked to machine learning model through an endpoint of web application.
● Though the Chrome browser enables safe browsing by default but that is not sufficient against exponentially growing phishing sites. It is proved that this Extension is efficient than Google safe browsing. Being prepared with Safe Browsing Extension can help world deal with cybercriminals better.
● An endpoint to book a ticket using a user’s name, phone number, and timings.
● An endpoint to update a ticket timing.
● An endpoint to view all the tickets for a particular time.
● An endpoint to delete a particular ticket.
● An endpoint to view the user’s details based on the ticket id.
● Mark a ticket as expired if there is a difference of 8 hours between the ticket timing
and current time.
● All the tickets which are expired automatically deleted.
● For a particular timing, a maximum of 20 tickets can be booked.
● REST paradigm implemented.
● SQLite database used.
● POSTMAN snaps of all APIs attached in README.
A Deep Learning Classifier to classify various species of Monkeys. ResNet architecture is used here as it tackles the degradation problem most common in deep networks, where the model accuracy gets saturated and then degrades rapidly. Confusion Matrix : The diagonal elements represent the number of images for which the predicted label is equal to the true label, while off-diagonal elements are those that are mislabeled by the classifier.
A tool that detects whether URL is legitimate or malicious using
machine learning. Integrated with Google Chrome Extension to
detect whether the page is safe or not. Implemented with an API
that returns status of safe or unsafe site.
Detects stance for controversial topic "Gun-laws" in USA, using
Data Science, Machine Learning and Deep Learning Models. It prompts for a
tweet from a user and returns appropriate stance, along with detailed Data Analysis.
This is a YouTube Content Analysis application that prompts for
a YouTube video URL and returns live content analysis of that
respective like number of hate speech or offensive language
used in that video.
It is a Reddit Flair Detector that takes a dynamic website as an
input and predicts 'flair' for the given URL based upon its title,
top comments of other users and URL based upon a trained
machine learning model.
Working on a data-set of past 10 years for minimum, maximum
temperatures and data-set for rainfall to study average rainfall
patterns and their dependence on minimum and maximum
temperatures in Ann Arbor, Michigan, United States.