Breakthroughs in Blockchain Technology: Functional Neural Networks Security Model for Permissionless Proof-of-Stake Blockchains against Benign Nodes
Tamara Saad Mohamed
Source abstract
Permissionless Proof-of-Stake (PoS) blockchain networks must demonstrate security and dependability as blockchain technology evolves. This research analyzes the ongoing need for a way to identify and eradicate rogue nodes inside such networks, also to advocate for the real-time rogue node’s identification in PoS blockchain networks by the application of neural network methodologies, particularly random neural networks. Experimental results indicate the ability of the developed model in differentiating legitimate blockchain nodes from malicious ones. The dataset in this paper is divided into two groups- malicious (Permissionless Proof-Of-Stake Blockchains) and non-malicious; this dataset is crucial for anyone interested in blockchain. The dataset includes details of the creation and validation of the PoS blocks. The paper aims to detect malicious nodes, analyze node behavior, and enhance security on PoS permissionless blockchains through data visualization. This paper shows the performance and process of the random neural network to refine and learn and then recognize the permissionless blockchain (malicious nodes) from the dataset of Proof-of-Stake Blockchain. We selected 100 records from the original dataset to examine them with our proposal. After analyzing the results, we found clearly how the proposal algorithm works properly with the proposed dataset to achieve fine accuracy and efficiency in the work to distinguish benign nodes from malicious (permission-less blockchain) nodes. This is the result of refining and learning the dataset using the random neural network for 340 instances, coming from the neural learning of 100 instances and 21 variables: [15 features: BlockHeight, UnixTimestamp, TxnFee (ETH), Block Generation Rate, TxnFee (Binary), Status (Tags), Stake Reward, Txnsize, Coin Days, Coin Age, Coin Stake, Stake Distribution Rate, Block Density, Block Score, Coin Day Weight) + 5 meta (node label, neural network, neural network0, neural network1, fold, selected) + no missing value), by two attributes: (node label, block score), two classes(0 non-malicious nodes,1 malicious nodes).
Evidence graph
No public relationships recorded yet.
Integrity note: This page is a factual metadata record created by deterministic ingestion. It is not a claim that the work moves a mathematical frontier or has been independently verified.