Hard Disk Drives (HDD) failure prediction is a challenging topic that has attracted much attention in recent years. Predicting failures HDD may avoid losing data thus improving reliability. Previous works on are based parametric approaches model healthy drives with Gaussian distribution. Although they achieved good results, the Gaussianity assumption not hold true. The following work proposes method for fault detection Mixture Model. A self-monitoring, analysis, and reporting technology...

Many complex natural and technological systems are commonly modeled as count-weighted graphs, where nodes represent entities, edges model relationships between them, edge weights define some counting statistics associated with each relationship. As graph data usually contain sensitive information about preserving privacy when releasing this type of becomes an important issue. In context, differential (DP) has become the de facto standard for release under strong guarantees. When dealing DP...

This work proposes many contributions to privacy in complex systems, mainly ones modeled as count-weighted graphs. As graph data usually contain users’ sensitive information, preserving when releasing this type of becomes a crucial issue. In context, differential (DP) has become the de facto standard for release under strong mathematical guarantees. However, various challenges persist effectively implementing DP data, including balancing protection with utility and scalability concerns. To...

As the amount of collected social network information in RDF format grows, development solutions for privacy individuals, their attributes and relationships with others becomes an important subject study. However, data are not well suitable this specific type data, mainly because they usually do consider between which crucial to semantic networks. Differential is one most techniques statistical queries and, although it has been extensively studied many papers, there still much research be...

Advancements in mobile computing techniques along with the pervasiveness of location-based services have generated a great amount trajectory data. These data can be used for various analysis purposes such as traffic flow analysis, infrastructure planning and understanding human behavior. However, publishing this may lead to serious risks privacy breach. Quasi-identifiers are points that linked external information identify individuals associated trajectories. Therefore, by analyzing...

Advancements in electronic fabrication technologies have facilitated the large-scale production of computer components, which are prone to faults over time. Despite availability fault-reporting tools provided by hardware manufacturers, there is a significant gap effectively utilizing textual reports due data scarcity. In this paper, we introduce FACTO dataset, comprehensive collection user on faulty components such as video cards, storage devices, motherboards, memory, and others. Data was...

With the increasing concerns over data privacy, preserving privacy of individuals in social network analysis has become crucial. This tutorial provides a comprehensive overview methods and techniques to protect individual while conducting analysis. We perform deep differential which is rigorous mathematical framework enabling accurate structure characteristics. Additionally, this explores variety examples case studies demonstrate application these practical scenarios.

Differential privacy is the state-of-the-art formal definition for data release under strong guarantees. A variety of mechanisms have been proposed in literature releasing noisy output numeric queries (e.g., using Laplace mechanism), based on notions global sensitivity and local sensitivity. However, although there has some work generic non-numeric Exponential lacks to reduce noise query output. In this work, we remedy shortcoming present dampening mechanism. We adapt notion setting leverage...

Differential privacy is the state-of-the-art formal definition for data release under strong guarantees. A variety of mechanisms have been proposed in literature releasing output numeric queries (e.g., Laplace mechanism and smooth sensitivity mechanism). Those guarantee differential by adding noise to true query's output. The amount added calibrated notions global local query that measure impact addition or removal an individual on Mechanisms use add less and, consequently, a more accurate...

Differential privacy is a model which gives strong guarantees. It was designed to make difficult distinguish individuals' records on statistical databases while maximizing data utility. approaches usually assume that database are sampled independently, i.e., each record of this independent the rest. However, assumption not always true in context real-world applications. In paper we propose DiPCoDing, novel approach calculate correlation between using clusterization. For matter, have...

Being able to detect faults in Hard Disk Drives (HDD) can lead significant benefits computer manufacturers, users and storage system providers. As a consequence, several works have focused on the development of fault detection algorithms for HDDs. Recently, promising results were achieved by methods using SMART (Self-Monitoring Analysis Reporting Technology) features anomaly based Mahalanobis distance. Nevertheless, performance such be seriously degraded when normality assumption data does...

Dados de tráfego redes são úteis para uma variedade aplicações. Geralmente as entidades que coletam esse tipo dado, por exemplo provedores internet (ISPs), compartilham suas informações rede com externas. Contudo, compartilhamento pode levar a violações privacidade dos indivíduos contidos nesses dados. Este trabalho propõe nova abordagem dados utilizando diferencial, um método tem como objetivo adicionar ruído sobre os originais. Resultados experimentais mostram proposta introduz menos nos...

The increasing adoption of solid-state drives (SSDs) due to their high performance and reliability has made failure prediction crucial for ensuring data integrity availability. Self-monitoring, Analysis, Reporting Technology (SMART) is a system that periodically reports various operational parameters facilitate early detection potential issues. Although many studies have used SMART attributes approaching this matter – as binary problem we test new ways predicting SSD failures, considering...

The correct functioning of Dynamic Random Access Memory (DRAM) is fundamental relevance to the servers in data centers. Therefore, being able detect server failure caused by memory errors development prediction methods that can be used avoid errors. Thus, ensuring continuous availability hosted services. In recent years, many authors proposed machine learning-based predict based on occurrence DRAM However, from previous works, one notice this a challenging task due lack and irregularity...

Differentially private selection mechanisms are fundamental building blocks for privacy-preserving data analysis. While numerous exist single-objective selection, many real-world applications require optimizing multiple competing objectives simultaneously. We present two novel differentially multi-objective selection: PrivPareto and PrivAgg. uses a Pareto score to identify solutions near the frontier, while PrivAgg enables weighted aggregation of objectives. Both support global local...

Serviços baseados em localização têm sido integrados às atividades diárias das pessoas. Entretanto, alguns desses serviços podem não ser confiáveis e levar a sérios riscos de violação privacidade. Este trabalho propõe uma nova técnica preservação privacidade dados, denominada PrivLBS, capaz assegurar que as localizações dos indivíduos serão facilmente reidentificadas por mal intencionados. Resultados avaliação experimental demonstram que, para ataques distância euclidiana, probabilidade...