Source Detection with Human Behavioral Dynamics
Budget: ₹600 – ₹1,500 INR
Meaning of the Topic
This project focuses on identifying the original source of infection or information spread in a network while considering human behavioral patterns.
In real-world networks, people do not share or transmit information uniformly. Their behavior affects how the spread happens. For example:
Panic sharing – People quickly share information during emergencies.
Selective forwarding – Users share information only with trusted contacts.
Delayed forwarding – Some users take time before forwarding messages.
Traditional source detection models assume that every user spreads information with the same probability. This project aims to improve source detection by incorporating human behavioral dynamics into the spreading model.
Approach
Create or simulate a network of users (nodes) and connections (edges).
Assign behavioral attributes to each user (panic level, trust score, forwarding probability).
Simulate infection or rumor spreading using these behavior parameters.
Train a machine learning model (such as a Graph Neural Network) to detect the original source.
Compare the proposed model with traditional source detection methods.
Novelty
The novelty of this project is the integration of human behavioral factors into the infection/rumor propagation model, which is usually ignored in traditional source detection approaches.
Expected Outcome
The system should be able to identify the most likely origin of infection or rumor spread in a network more accurately by considering human behavior patterns.
This project focuses on identifying the original source of infection or information spread in a network while considering human behavioral patterns.
In real-world networks, people do not share or transmit information uniformly. Their behavior affects how the spread happens. For example:
Panic sharing – People quickly share information during emergencies.
Selective forwarding – Users share information only with trusted contacts.
Delayed forwarding – Some users take time before forwarding messages.
Traditional source detection models assume that every user spreads information with the same probability. This project aims to improve source detection by incorporating human behavioral dynamics into the spreading model.
Approach
Create or simulate a network of users (nodes) and connections (edges).
Assign behavioral attributes to each user (panic level, trust score, forwarding probability).
Simulate infection or rumor spreading using these behavior parameters.
Train a machine learning model (such as a Graph Neural Network) to detect the original source.
Compare the proposed model with traditional source detection methods.
Novelty
The novelty of this project is the integration of human behavioral factors into the infection/rumor propagation model, which is usually ignored in traditional source detection approaches.
Expected Outcome
The system should be able to identify the most likely origin of infection or rumor spread in a network more accurately by considering human behavior patterns.