The problem Phishing Net solves

"Dear Customer, you have won a cash prize of... Please click on this link to..."

How many of you have seen these kinds of messages?

In light of the expanding fraudulent exercises, we have developed Phishing Net, a one-stop solution to identify and report scams and safeguard public interests.

Phishing Net is a flutter application that classifies the received messages into safe, unsafe, and spam messages.

Phishing Net is a small step towards solving the 16th Sustainable Development Goal set by the United Nations that calls for setting up effective, accountable, and transparent institutions at all levels.

Challenges we ran into

Some of the significant challenges we faced building the project were:

• Dealing with False Positives in the dataset: We experimented with various ML algorithms to finally settle for XGBoost at an accuracy of 94%. We plan on improving it further by increasing our dataset.
• Connecting our ML model with our flutter application: We were able to achieve this using REST-API.
• User SMS Database Management System: Maintaining the model for the User SMS initially posed a problem before and after classification but we were able to bridge that after hours of brainstorming.