H2E: A Privacy Provisioning Framework for Collaborative Filtering Recommender System

Abstract

A Recommender System (RS) is the most significant technologies that handle the information overload problem of Retrieval Information by suggesting users with correct and related items. Today, abundant recommender systems have been developed for different fields and we put an effort on collaborative filtering (CF) recommender system. There are several problems in the recommender system such as Cold Start, Synonymy, Shilling Attacks, Privacy, Limited Content Analysis and Overspecialization, Grey Sheep, Sparsity, Scalability and Latency Problem. The current research explored the privacy in CF recommender system and defined the perspective privacy attributes (user's identity, password, address, and postcode/location) which are required to be addressed. Using the base models as Homomorphic and Hash Encryption scheme, we have proposed a hybrid model Homomorphic Hash Encryption (H2E) model that addressed the privacy issues according to defined objectives in the current study. Furthermore, in order to evaluate the privacy level, H2E was implementing in medicine recommender system and compared the consequences with existing state-of-the-art privacy protection mechanisms. It was observed that H2E outperform to other models with respect to determined privacy objectives. Leading to user's privacy, H2E can be considered a promising model for CF recommender systems.

COVID Detection Using Chest X-ray Images & Deep Learning CNN Mechanism

“…Besides, when CT is stood out from Xray, the proportion of radiation got by the patient is more. In this way, it is recommended to use a profound learning model with Xray imagery, as it is continuously open with lower radiation segment when stood out from CT [60][61][62][63][64][65]. Patients examined as Coronavirus +ve by the model can be facilitated to state of the art natural surroundings for insistence, followed by treatment right away.…”

“…And we witness that morethe model is deeper network, its accuracy will be higher. As the ResNet152 has much higher accuracy than ResNet101 [65][66][67][68][69][70][71][72]. Even though the depth has increased significantly the ResNet152 layer architectures having 11.4 billion FLOPS (floating point operations per second) which has minimum complexity than other models.…”

A Secured Intrusion Detection System for Mobile Edge Computing

“…Malware Injection Attack: A malware injection attack [29][30][31][32], also known as a code injection attack, is a type of cybersecurity threat where malicious code is inserted into a legitimate application or system with the intent of compromising its functionality, stealing sensitive information, or gaining unauthorized access. This form of attack exploits vulnerabilities in software, allowing the attacker to execute arbitrary code and manipulate the targeted system for malicious purposes.…”

Theoretical Studies of the Effect of Loads on Aircraft Disc Brakes

“…Для описания процесса торможения на примере фрикционного элемента рассмотрим нагревание постоянным тепловым потоком q, который описывается краевой задачей (1) с граничными условиями второго рода [7,8]:…”