Providing Highest Privacy Preservation Scenario for Achieving Privacy in Confidential Data
Journal Title: International Journal of Experimental Research and Review - Year 2024, Vol 39, Issue 3
Abstract
Machine learning algorithms have been extensively employed in multiple domains, presenting an opportunity to enable privacy. However, their effectiveness is dependent on enormous data volumes and high computational resources, usually available online. It entails personal and private data like mobile telephone numbers, identification numbers, and medical histories. Developing efficient and economical techniques to protect this private data is critical. In this context, the current research suggests a novel way to accomplish this, combining modified differential privacy with a more complicated machine learning (ML) model. It is possible to assess the privacy-specific characteristics of single or multiple-level models using the suggested method, as demonstrated by this work. It then employs the gradient values from the stochastic gradient descent algorithm to determine the scale of Gaussian noise, thereby preserving sensitive information within the data. The experimental results show that by fine-tuning the parameters of the modified differential privacy model based on the varied degrees of private information in the data, our suggested model outperforms existing methods in terms of accuracy, efficiency and privacy.
Authors and Affiliations
Pinkal Jain, Vikas Thada, Deepak Motwani
Keywords
Related Articles
- EP ID EP735998
- DOI 10.52756/ijerr.2024.v39spl.015
- Views 37
- Downloads 0
Agrometeorological indices, physiological growth parameters and performance of finger millet as influenced by different cultivars under hot and sub humid region of Odisha
Present-day agriculture is under tremendous pressure due to various factors such as climate change, degradation of soil and water quality, and yield plateauing of major food crops. Under these scenarios, there is an urge...
Click Chemistry: copper, ruthenium catalyzed and photoinduced
Click chemistry is an extremely powerful method for covalent conjugation of molecular entities quickly and efficiently. Click chemistry knitted the threads between two different molecular entities that have created inter...
Design, Performance and Economic evaluation of a 4kW Grid-interactive Solar PV Rooftop in Odisha using Pvsyst
The global decline of fossil fuels has required the development of alternative energy sources to satisfy the electricity demand. This article assesses and inaugurated the viability of grid-interactive photovoltaic system...
Review on Phyto-chemistry & pharmacological activity of Melia azedarach
Medicinal plants as essential additives in foods, beverages, and medications have been a part of human culture since the dawn of time. Medicinal plants' nutritional, pharmacological, biological, and toxicological charact...
Luminescent metal-organic frameworks
Metal-organic frameworks (MOFs) show a wide range of luminescent behaviours. In this review, there is a brief discussion on the origins of MOF luminosity due to various type of linkers and their special arrangement in MO...