Fig 2 - uploaded by Ezgi Bingöl
Content may be subject to copyright.
Context in source publication
Similar publications
ATM card fraud is increasing gradually with the expansion of modern technology and global communication. In the whole world, it is resulting in the loss of billions of dollars each year. Fraud detection systems have become essential for all ATM card issuing banks to minimize their losses. The main goals are, firstly, to review alternative technique...
One of the challenging issues in user‐click data of online advertising is the uneven class distribution which biases classification models. Resampling the data is a popular choice for obtaining class balance. However, oversampling results in overfitting, whilst under‐sampling results in information loss. Moreover, enhancing separability between sam...
![Figure 1: The revised taxonomy of anomaly types by Lai et al. [14]](publication/375784220/figure/fig1/AS:11431281206192000@1700578220026/The-revised-taxonomy-of-anomaly-types-by-Lai-et-al-14_Q320.jpg)
![Figure 2: An illustration (borrowed from Darban et al. [15]) of the...](publication/375784220/figure/fig2/AS:11431281206164512@1700578220151/An-illustration-borrowed-from-Darban-et-al-15-of-the-taxonomy-of-anomaly-types_Q320.jpg)
Anomaly detection has a wide variety of applications, ranging from intrusion detection in cybersecurity to fraud detection in finance. Among the most prominent applications is predictive maintenance in manufacturing, which involves performing maintenance only when truly necessary, based on the condition of relevant equipment instead of following a...
Our research addresses the actual behavior of the credit-card fraud detection environment where financial transactions containing sensitive data must not be amassed in a considerable amount to train robust classifiers. We introduce an adaptive learning approach that adjusts frequently and efficiently to new transaction chunks; each chunk is discard...
Citations
... The search identified 34 interventions ( Table 4) that aim to detect anomalies, threats, or unwanted behaviour. While some studies analysed human behaviour, facial expressions, or lipmovement to identify threats in individual people (Anagnostopoulos, 2014;Byun, Nasridinov, & Park, 2014;Rothkrantz, 2017b;Sajjad et al., 2018), others sought to detect fraudulent behaviour through the analysis of big data and crowd movement patterns (Cemgil, Kurutmaz, Cezayirli, Bingol, & Sener, 2017;Gupta, Chakraborty, & Mondal, 2017;Liu, Ni, & Krishnan, 2014;Rocher, Taha, Parra, & Lloret, 2018;Sadgali, Sael, & Benabbou, 2018). Even though many of these interventions operated to a large extent the sensor layer of the smart city and relied on already existent actuators, they often did include secondary functions. ...
The implementation of smart technology in cities is often hailed as the solution to many urban challenges such as transportation, waste management, and environmental protection. Issues of security and crime prevention, however, are in many cases neglected. Moreover, when researchers do introduce new smart security technologies, they rarely discuss their implementation or question how new smart city security might affect traditional policing and urban planning processes. This systematic review explores the recent literature concerned with new ‘smart city’ security technologies and aims to investigate to what extent these new interventions correspond with traditional functions of security interventions. Through an extensive literature search we compiled a list of security interventions for smart cities and suggest several changes to the conceptual status quo in the field. Ultimately, we propose three clear categories to categorise security interventions in smart cities: Those interventions that use new sensors but traditional actuators, those that seek to make old systems smart, and those that introduce entirely new functions. These themes are then discussed in detail and the importance of each group of interventions for the overall field of urban security and governance is assessed.
... Mesmo no cenário de Smart Grids, existe a possibilidade de ocorrerem problemas de perda ou leitura incorreta dos dados nos Smart Meters devido a fraudes ou falhas no processo de transmissão e medição, de maneira que dados corrompidos ou faltantes são medidos e transmitidosà central gerando inconsistência [Cemgil et al. 2017] [Chen et al. 2010]. Tais problemas não são enfrentados apenas nessaárea, podendo ser também um empecilho na geofísica, como relatado por [Li et al. 2017], onde a falta de dadosé um fenômeno comum nas medidas do ambiente espacial. ...
... Utilizando dados reais das curvas de carga da British Columbia Transmission Corporation (BCTC), o experimento e a avaliação foram realizados, mostrando a eficácia da solução apresentada. [Cemgil et al. 2017] assim como em [Chen et al. 2010] trata do problema de dados faltantes e atípicos em um AMR, e mostra que os dados de consumo de energia elétrica carregam boas indicações de fraude no sistema, caso estas estejam presentes. Os autores criaram dois algoritmos para poder interpolar dados faltantes e, dessa forma, poder detectar fraudes, denominados Auto-Regressive (AR) e Non-negative Matrix Factorization (NMF). ...
Abstract: Smart Grids are networks that are responsable for the energy distribuition on a safe way and promote a fair measure of consumption, for having a big
quantity of sensors to monitor and record diferente quantity of data throughout the day, they may fail to collect informations, producing missing data or
invalid, affecting the quality of the service. Therefore, this article presents the proposal of an adaptive algorithm, built from the performance evaluation of
two algorithms used for the imputation of missing data, Spline and Singular Spectrum Analysis- SSA. Performance evaluation shows significant improvements in the imputation of missing data using the built algorithm, allowing a more accurate measurement of consumption even with missing data.
Resumo: As Smart Grids são redes responsáveis por distribuir a energia de forma segura e promover uma medição justa do consumo. Por terem uma grande quantidade de sensores para monitorar e registrar diferentes quantidades de dados ao longo do dia, podem deixar de coletar informações, produzindo dados ausentes ou inválidos, afetando a qualidade do serviço. Esse artigo apresenta a proposta de um algoritmo adaptativo, construído a partir da avaliação de desempenho de dois algoritmos utilizados para a imputação de dados faltantes, o Spline e o Singular Spectrum Analysis(SSA). A avaliação de desempenho mostra melhorias significativas na imputação de dados faltantes com o algoritmo construído, permitindo uma medição mais precisa do consumo mesmo com dados faltantes.
... An approach for implementing two communication media, the Power Line Carrier (PLC) and the General Packet Radio Service (GPRS) in the Automated Meter Reading (AMR) and the Advanced Metering Infrastructure (AMI) project is explained in [15]. Cemgil et al. [16] planned a mathematical model to be able to interpolate missing data and detect fraud of electricity meters that occur due to communication faults, device faults, or energy fraud. ...
Even in this technological era where almost every field of life is automated and digitized, the arena concerning electricity meter reading and billing in the perspective of Bangladesh still lies within the range of manual processes and thus requires additional human labour along with extra overhead. Alongside due to lack of proper availability of usage statistics accessible from user end, the wastage of electricity is on a rise. Electricity being a dominant source of energy, its wastage occurring every day in every possible way should be avoided and dealt with proper concern. This paper focuses on creating an automated bill monitoring system along with a mobile application that will automate the meter reading procedure by collecting the consumed electrical energy of houses. The system will also encourage the users to reduce wastage of electricity which is an essential part of our everyday life. A prototype has been developed and evaluated in an academic environment to assess the system to compare it with the existing meter.
... In the real time traffic flow detection was shown that, the ability of some system can be enhance in adapting the changes of traffic patterns by treating the outlier before the forecasting was applied. The detected outliers were recommended and should be necessary to be treated in performing the short term traffic condition forecasting since it might indicate the changes in measurement errors, pattern changes and parameters [7]. Some researchers prefer to apply robust method as outlier treatment. ...
