Anomaly Detection in Social Media Using Recurrent Neural Network

Shah, Shamoz; Goyal, Madhu

doi:10.1007/978-3-030-22747-0_6

Cited by 3 publications

(3 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…According to [14], VAD systems are either manually built by experts setting thresholds on data or constructed automatically by learning from the available data through Machine Learning (ML). VAD is widely used in many applications such as fraud detection [15], [16], image processing [17], [18], sensor networks [19], [20], medical health [21], [22], intrusion detection [23], IT security [24], [25], [26], and social media [27], [28]. Fig.…”

Section: Video Anomaly Detectionmentioning

confidence: 99%

A Review on Video Anomaly Detection Datasets

Yossef,

Gamal,

Abdel-Kader

et al. 2023

Suez Canal Engineering, Energy and Environmental Science

View full text Add to dashboard Cite

In recent years, Video Anomaly Detection (VAD) has received a lot of attention and has become a popular research topic. This is due to their immense potential in a variety of fields, including healthcare monitoring, surveillance/crowd analysis, sports, Ambient Assistive Living (AAL), event analysis, and security. Manually detecting and analysing improper behavior was a hard process, particularly in real-time scenarios, resulting in a high demand for smart surveillance systems. Moreover, the availability of data plays a vital role in training and evaluating models. Datasets in VAD are typically composed of sequences of frames or videos, some of which depict normal activities and others that depict anomalous or unusual events. These datasets provide a rich resource that encapsulates everyday routine actions alongside irregular or unusual events, fostering the development and assessment of robust anomaly detection models. This paper provides an extensive review of the most popular and recent datasets in VAD including an extensive comparison between them.

show abstract

Section: Video Anomaly Detectionmentioning

confidence: 99%

A Review on Video Anomaly Detection Datasets

Yossef,

Gamal,

Abdel-Kader

et al. 2023

Suez Canal Engineering, Energy and Environmental Science

View full text Add to dashboard Cite

show abstract

“…Application of traditional approaches in combination with deep learning has also been studied. For example, the authors in [32] have demonstrated that before training a RNN model, an initial clustering with K-NN can enhance the detection of outliers in social media. A more recent survey paper that brings together previous approaches on deep learning in the domain of wireless and mobile networking domain is given in [33].…”

Section: Deep Learning Based Ad Approachesmentioning

confidence: 99%

Detecting Mobile Traffic Anomalies Through Physical Control Channel Fingerprinting: A Deep Semi-Supervised Approach

et al. 2019

View full text Add to dashboard Cite

Among the smart capabilities promised by the next generation cellular networks (5G and beyond), it is fundamental that potential network anomalies are detected and timely treated to avoid critical issues concerning network performance, security, public safety. In this paper, we propose a comprehensive framework for detecting network anomalies using mobile traffic data: collecting data from the LTE Physical Downlink Control Channel (PDCCH) of different eNodeBs, we implement deep learning algorithms in a semi-supervised way to detect potential traffic anomalies that are generated, for example, by unexpected crowd gathering. With respect to other types of mobile dataset, using LTE PDCCH information, we are able to obtain fine-grained and high-resolution data for the users that are connected to the LTE eNodeB. Through a semi-supervised approach, algorithms are trained to detect anomalies using only one class of traffic samples. We design two algorithms based on stacked-LSTM Neural Networks: 1) LSTM Autoencoder (LSTM-AE), in which the objective is to reconstruct the traffic samples 2) LSTM traffic predictor (LSTM-PRED), where the goal is to predict the traffic in the next time-instants, based on historical data. In both cases, we analyze the reconstruction (or prediction) error to assess if the mobile traffic presents anomalies or not. Using the F1-score as metric, we demonstrate that the proposed methods are able to identify the anomalous traffic periods, beating a benchmark that comprises different state-of-the-art algorithms for anomaly detection.

show abstract

“…Application of traditional approaches in combination with deep learning have also been studied. For example, the authors in [74] [76,77].…”

Section: Deep Learning Based Ad Approachesmentioning

confidence: 99%

Data analytics for mobile traffic in 5G networks using machine learning techniques

Trinh

View full text Add to dashboard Cite

This thesis collects the research works I pursued as Ph.D. candidate at the Universitat Politecnica de Catalunya (UPC). Most of the work has been accomplished at the Mobile Network Department Centre Tecnologic de Telecomunicacions de Catalunya (CTTC). The main topic of my research is the study of mobile network traffic through the analysis of operative networks dataset using machine learning techniques. Understanding first the actual network deployments is fundamental for next-generation network (5G) for improving the performance and Quality of Service (QoS) of the users. The work starts from the collection of a novel type of dataset, using an over-the-air monitoring tool, that allows to extract the control information from the radio-link channel, without harming the users’ identities. The subsequent analysis comprehends a statistical characterization of the traffic and the derivation of prediction models for the network traffic. A wide group of algorithms are implemented and compared, in order to identify the highest performances. Moreover, the thesis addresses a set of applications in the context mobile networks that are prerogatives in the future mobile networks. This includes the detection of urban anomalies, the user classification based on the demanded network services, the design of a proactive wake-up scheme for efficient-energy devices. Esta tesis recoge los trabajos de investigación que realicé como Ph.D. candidato a la Universitat Politecnica de Catalunya (UPC). La mayor parte del trabajo se ha realizado en el Centro Tecnológico de Telecomunicaciones de Catalunya (CTTC) del Departamento de Redes Móviles. El tema principal de mi investigación es el estudio del tráfico de la red móvil a través del análisis del conjunto de datos de redes operativas utilizando técnicas de aprendizaje automático. Comprender primero las implementaciones de red reales es fundamental para la red de próxima generación (5G) para mejorar el rendimiento y la calidad de servicio (QoS) de los usuarios. El trabajo comienza con la recopilación de un nuevo tipo de conjunto de datos, utilizando una herramienta de monitoreo por aire, que permite extraer la información de control del canal de radioenlace, sin dañar las identidades de los usuarios. El análisis posterior comprende una caracterización estadística del tráfico y la derivación de modelos de predicción para el tráfico de red. Se implementa y compara un amplio grupo de algoritmos para identificar los rendimientos más altos. Además, la tesis aborda un conjunto de aplicaciones en el contexto de redes móviles que son prerrogativas en las redes móviles futuras. Esto incluye la detección de anomalías urbanas, la clasificación de usuarios basada en los servicios de red demandados, el diseño de un esquema de activación proactiva para dispositivos de energía eficiente.

show abstract

Anomaly Detection in Social Media Using Recurrent Neural Network

Cited by 3 publications

References 10 publications

A Review on Video Anomaly Detection Datasets

A Review on Video Anomaly Detection Datasets

Detecting Mobile Traffic Anomalies Through Physical Control Channel Fingerprinting: A Deep Semi-Supervised Approach

Data analytics for mobile traffic in 5G networks using machine learning techniques

Contact Info

Product

Resources

About