Abstract:
"When we consider election systems or voting systems ensuring secure and accurate data
processing is paramount. The Private data of the general public could potentially be used ny political parties for doing various analysis without consent. The integrity and openness of
traditional procedures are frequently compromised when they are used across several voting locations. To overcome these obstacles, this study suggests the DFVote System which is
based on blockchain technology. DFVote makes use of blockchain technology to safely
handle and verify model parameters from numerous voting stations dispersed over different areas. Using station-specific data, each station independently trains local models, which are then combined and improved via Federated Learning methods.
Important parts consist of an InterPlanetary File System (IPFS) integration for decentralized model parameter storage, a Ganache Blockchain for demonstration purposes, and a React-
based interface for data input and display. With the use of tamper-proof audit trails made
possible by the Blockchain, the system records IPFS hashes to guarantee data integrity and transparency. This method not only makes election data more secure and reliable, but it also makes it possible to train and validate models efficiently in dispersed situations. A thorough
assessment of vote analysis models is used to show how successful this is and validates its applicability in actual election situations.
"
