Author(s):
1. DEEPTI JOON:
GD Goenka University, Gurugram, India
2. KHYATI CHOPRA:
School of Engineering Sciences & Technology, Jamia Hamdard, India
Abstract:
The Cyber-Physical System (CPS) utilizes Learning Enabled Components (LECs) with neural networks for understanding and decision-making tasks. Neural Networks are commonly used for reasoning and making predictions about energy forecasts, but subsequently, the prediction-based application in security basic frameworks is not effective. LECs can work simpler as CPS if their expectations could be supplemented with a proper certainty in resource measurement that evaluates the amount of output. The paper presents a methodology which is Long Short Term Memory (LSTM)-Adam optimizer based Inductive Conformal Prediction (ICP) for proper usage of resources. The Triplet Network is designed to learn the input data information for estimating the comparability among the test models collects the information. The main aim of the research is to perform forecasting certainty to improve the learning of the neural network classifier. The present approach will select a significant level for overcoming the trade-off error and the main aim is to reduce the false alarms. The present research performs multiple predictions for the Triplet with k-Nearest Neighbour (k-NN) for Non-Conformity Measure (NCM) function that shows significant improvement at a higher level as LSTM and showed lesser error values of 5.86 when compared with the existing K-NN based ICP model that obtained error values of 16.5 and Triplet K-NN of 9.2.
Page(s):
621-629
DOI:
DOI not available
Published:
Journal: Journal of Theoretical and Applied Information Technology, Volume: 100, Issue: 3, Year: 2022
Keywords:
Cyberphysical systems
,
Inductive Conformal Prediction
,
Long Short Term MemoryAdam Optimizer
,
kNearest Neighbour
,
Triplet
References:
References are not available for this document.
Citations
Citations are not available for this document.