Search results

Organizations in many domains generate a considerable amount of heterogeneous data every day. Such data can be processed to enhance these organizations’ decisions in real time. However, storing and processing large and varied datasets (known as big data) is challenging to do in real time. In machine learning, streaming feature selection has always been considered a superior technique for selecting the relevant subset features from highly dimensional data and thus reducing learning complexity. In the relevant literature, streaming feature selection refers to the features that arrive consecutively over time; despite a lack of exact figure on the number of features, numbers of instances are well-established. Many scholars in the field have proposed streaming-feature-selection algorithms in attempts to find the proper solution to this problem. This paper presents an exhaustive and methodological introduction of these techniques. This study provides a review of the traditional feature-selection algorithms and then scrutinizes the current algorithms that use streaming feature selection to determine their strengths and weaknesses. The survey also sheds light on the ongoing challenges in big-data research.

Streaming audio/video contents over the Internet requires large network bandwidth and timely delivery of media data. A streaming session is generally long and also needs a large I/O bandwidth at the streaming server. A streaming server, however, has limited network and I/O bandwidth. For this reason, a streaming server alone cannot scale a streaming service well. An entire audio/video media file often cannot be cached due to intellectual property right concerns of the content owners, security reasons, and also due to its large size. This makes a streaming service hard to scale using conventional proxy servers. Media file compression using variable‐bit‐rate (VBR) encoding is necessary to get constant quality video playback although it produces traffic bursts. Traffic bursts either waste network bandwidth or cause hiccups in the playback. Large network latency and jitter also cause long start‐up delay and unwanted pauses in the playback, respectively. In this paper, we propose a proxy based constant‐bit‐rate (CBR)‐transmission scheme for VBR‐encoded videos and a scalable streaming scheme that uses a CBRtransmission scheme to stream stored videos over the Internet. Our CBR‐streaming scheme allows a server to transmit a VBRencoded video at a constant bit rate, close to its mean encoding bit rate, and deals with the network latency and jitter issues efficiently in order to provide quick and hiccup free playback without caching an entire media file. Our scalable streaming scheme also allows many clients to share a server stream. We use prefix buffers at the proxy to cache the prefixes of popular videos, to minimize the start‐up delay and to enable near mean bit rate streaming from the server as well as from the proxy. We use smoothing buffers at the proxy not only to eliminate jitter and traffic burst effects but also to enable many clients to share the same server stream. We present simulation results to demonstrate the effectiveness of our streaming scheme.

Peer‐to‐peer (P2P) streaming quickly emerges as an important application over the internet. A lot of systems have been implemented to support peer‐to‐peer media streaming. However, some problems still exist. These problems include non‐guaranteed communication efficiency, limited upload capacity and dynamics of suppliers which are all related to the overlay topology design. The purpose of this paper is to propose a novel overlay construction framework for peer‐to‐peer streaming.

To exploit the bandwidth resource of neighboring peers with low communication delay, application of the grouping method was proposed to construct a flexible two‐layered locality‐aware overlay network. In the proposed overlay, peers are clustered into locality groups according to the communication delays of peers. These locality groups are interconnected with each other to form the top layer of the overlay. In each locality group, peers form an overlay mesh for transmitting stream to other peers of the same group. These overlay meshes form the bottom layer of the overlay.

Through simulations, the performance was compared in terms of communication efficiency, source‐to‐end delivery efficiency and reliability of the delivery paths of the proposed solution currently. Simulation results show that the proposed method can achieve the construction of a scalable, efficient and stable peer‐to‐peer streaming environment.

The new contributions in this paper are a novel framework which includes the adaptability, maintenance and optimization schemes to adjust the size of overlay dynamically according to the dynamics of peers; and considering the importance of locality of peers in the system.

Vehicle companion is one of the most common companion patterns in daily life, which has great value to accident investigation, group tracking, carpooling recommendation and road planning. Due to the complexity and large scale of vehicle sensor streaming data, existing work were difficult to ensure the efficiency and effectiveness of real-time vehicle companion discovery (VCD). This paper aims to provide a high-quality and low-cost method to discover vehicle companions in real time.

This paper provides a real-time VCD method based on pro-active data service collaboration. This study makes use of dynamic service collaboration to selectively process data produced by relative sensors, and relax the temporal and spatial constraints of vehicle companion pattern for discovering more potential companion vehicles.

Experiments based on real and simulated data show that the method can discover 67% more companion vehicles, with 62% less response time comparing with centralized method.

To reduce the amount of processing streaming data, this study provides a Service Collaboration-based Vehicle Companion Discovery method based on proactive data service model. And this study provides a new definition of vehicle companion through relaxing the temporal and spatial constraints for discover companion vehicles as many as possible.

The purpose of this paper is to conduct an in-depth exploration of the special context and user experiences of live video streaming and to provide insights regarding an interpretation of the contextualization experiences model.

This study used netnography, online interviews and the physical travel of researchers to the field for field participation and observations. The combination of netnography and online interviews combined online and offline studies to achieve greater consistency in the data collection, analysis and other processes.

The findings of the study can be classified into a three-stage situational context approach, which is presented in the form of propositions. Finally, the insights of the contextualization experiences model are presented.

This study resulted in the development of a substantive theory that provides insight into interpreting the contextualization experiences model. The theory was developed based on raw data to enable it to explain the phenomena in the context of similar instances of live video streaming.

This study aims to examine if the participation of live-stream influencers (LSIs) affects tipping frequency on live streaming platforms, and further investigate the mediating and moderating mechanisms.

Quasi-experiment and difference-in-differences models are used for data analysis. Propensity score matching is used to address potential unobservable endogeneity.

Real-time live streaming data reveal that LSIs’ participation significantly improves tipping frequency in live streaming rooms. Also, more users are attracted to the live streaming rooms and more users become active in participation. Additionally, the positive impact of LSIs’ participation is enhanced in the live streaming rooms with a greater number of relationship links between users.

The findings clarify the new role of influencers and reveal the mechanisms on how LSIs benefit the platforms.

The findings offer novel insights into implementing influencer marketing to interactive social media platforms, by encouraging influencer participation, user relationship building and influencer network growth.

This study highlights the value of LSIs for interactive social media platforms in terms of organic growth, revenue generation and cost reduction.

The purpose of this paper is to articulate whether consumers’ use of music via streaming service benefits niche products and diversified consumption of music. It examines does winner take all or is long tail achieved in the digital music market.

To investigate the degree of concentration in the digital music sales, this study measures multiple concentration metrics using the top 100 songs for 245 weeks listed on the Korean music ranking chart.

Conflicting results are found between the analyses based on short-run and long-run data. When sales distributions are compared weekly or monthly, the results show that streaming services have a less concentrated sales distribution than download services. However, the result becomes the opposite in the long-run analysis (i.e. one year).

This study proposes that the non-technological drivers such as the beneficial addiction of music consumption can be a crucial driver affecting the usage concentration in music industry, coupled with the royalty policy of access-based services.

The aim of continuous learning is to obtain and fine-tune information gradually without removing the already existing information. Many conventional approaches in streaming data classification assume that all arrived new data is completely labeled. To regularize Neural Networks (NNs) by merging side information like user-provided labels or pair-wise constraints, incremental semi-supervised learning models need to be introduced. However, they are hard to implement, specifically in non-stationary environments because of the efficiency and sensitivity of such algorithms to parameters. The periodic update and maintenance of the decision method is the significant challenge in incremental algorithms whenever the new data arrives.

Hence, this paper plans to develop the meta-learning model for handling continuous or streaming data. Initially, the data pertain to continuous behavior is gathered from diverse benchmark source. Further, the classification of the data is performed by the Recurrent Neural Network (RNN), in which testing weight is adjusted or optimized by the new meta-heuristic algorithm. Here, the weight is updated for reducing the error difference between the target and the measured data when new data is given for testing. The optimized weight updated testing is performed by evaluating the concept-drift and classification accuracy. The new continuous learning by RNN is accomplished by the improved Opposition-based Novel Updating Spotted Hyena Optimization (ONU-SHO). Finally, the experiments with different datasets show that the proposed learning is improved over the conventional models.

From the analysis, the accuracy of the ONU-SHO based RNN (ONU-SHO-RNN) was 10.1% advanced than Decision Tree (DT), 7.6% advanced than Naive Bayes (NB), 7.4% advanced than k-nearest neighbors (KNN), 2.5% advanced than Support Vector Machine (SVM) 9.3% advanced than NN, and 10.6% advanced than RNN. Hence, it is confirmed that the ONU-SHO algorithm is performing well for acquiring the best data stream classification.

This paper introduces a novel meta-learning model using Opposition-based Novel Updating Spotted Hyena Optimization (ONU-SHO)-based Recurrent Neural Network (RNN) for handling continuous or streaming data. This is the first work utilizes a novel meta-learning model using Opposition-based Novel Updating Spotted Hyena Optimization (ONU-SHO)-based Recurrent Neural Network (RNN) for handling continuous or streaming data.

A new research domain known as the Quantified Self has recently emerged and is described as gaining self-knowledge through using wearable technology to acquire information on self-monitoring activities and physical health related problems. However, very little is known about the impact of time window models on discovering self-quantified patterns that can yield new self-knowledge insights. This paper aims to discover the self-quantified patterns using multi-time window models.

This paper proposes a multi-time window analytical workflow developed to support the streaming k-means clustering algorithm, based on an online/offline approach that combines both sliding and damped time window models. An intervention experiment with 15 participants is used to gather Fitbit data logs and implement the proposed analytical workflow.

The clustering results reveal the impact of a time window model has on exploring the evolution of micro-clusters and the labelling of macro-clusters to accurately explain regular and irregular individual physical behaviour.

The preliminary results demonstrate the impact they have on finding meaningful patterns.

The purpose of this article is to propose and evaluate a novel system architecture for Smart City applications which uses ontology reasoning and a distributed stream processing framework on the cloud. In the domain of Smart City, often methodologies of semantic modeling and automated inference are applied. However, semantic models often face performance problems when applied in large scale.

The problem domain is addressed by using methods from Big Data processing in combination with semantic models. The architecture is designed in a way that for the Smart City model still traditional semantic models and rule engines can be used. However, sensor data occurring at such Smart Cities are pre-processed by a Big Data streaming platform to lower the workload to be processed by the rule engine.

By creating a real-world implementation of the proposed architecture and running simulations of Smart Cities of different sizes, on top of this implementation, the authors found that the combination of Big Data streaming platforms with semantic reasoning is a valid approach to the problem.

In this article, real-world sensor data from only two buildings were extrapolated for the simulations. Obviously, real-world scenarios will have a more complex set of sensor input values, which needs to be addressed in future work.

The simulations show that merely using a streaming platform as a buffer for sensor input values already increases the sensor data throughput and that by applying intelligent filtering in the streaming platform, the actual number of rule executions can be limited to a minimum.