Search results
1 – 10 of over 11000Noura AlNuaimi, Mohammad Mehedy Masud, Mohamed Adel Serhani and Nazar Zaki
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…
Abstract
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.
Details
Keywords
Renuka Devi D. and Sasikala S.
The purpose of this paper is to enhance the accuracy of classification of streaming big data sets with lesser processing time. This kind of social analytics would contribute to…
Abstract
Purpose
The purpose of this paper is to enhance the accuracy of classification of streaming big data sets with lesser processing time. This kind of social analytics would contribute to society with inferred decisions at a correct time. The work is intended for streaming nature of Twitter data sets.
Design/methodology/approach
It is a demanding task to analyse the increasing Twitter data by the conventional methods. The MapReduce (MR) is used for quickest analytics. The online feature selection (OFS) accelerated bat algorithm (ABA) and ensemble incremental deep multiple layer perceptron (EIDMLP) classifier is proposed for Feature Selection and classification. Three Twitter data sets under varied categories are investigated (product, service and emotions). The proposed model is compared with Particle Swarm Optimization, Accelerated Particle Swarm Optimization, accelerated simulated annealing and mutation operator (ASAMO). Feature Selection algorithms and classifiers such as Naïve Bayes, support vector machine, Hoeffding tree and fuzzy minimal consistent class subset coverage with the k-nearest neighbour (FMCCSC-KNN).
Findings
The proposed model is compared with PSO, APSO, ASAMO. Feature Selection algorithms, and classifiers such as Naïve Bayes (NB), support vector machine (SVM), Hoeffding Tree (HT), and Fuzzy Minimal Consistent Class Subset Coverage with the K-Nearest Neighbour (FMCCSC-KNN). The outcome of the work has achieved an accuracy of 99%, 99.48%, 98.9% for the given data sets with the processing time of 0.0034, 0.0024, 0.0053, seconds respectively.
Originality/value
A novel framework is proposed for Feature Selection and classification. The work is compared with the authors’ previously developed classifiers with other state-of-the-art Feature Selection and classification algorithms.
Details
Keywords
Futao Zhao and Zhong Yao
The purpose of this paper is to identify the impact factors that might influence audiences' voluntary donation to content creators on the online platforms, and to build an…
Abstract
Purpose
The purpose of this paper is to identify the impact factors that might influence audiences' voluntary donation to content creators on the online platforms, and to build an effective prediction model by considering both content and creator-related features.
Design/methodology/approach
This study collected the real-world data of content consumption from Xueqiu.com and extracted both content and creator characteristics from the data set. The best donation prediction model based on such features was determined by evaluating four prevalent classifiers with various performance metrics. Furthermore, three feature selection methods were applied to validate the robustness of the constructed model, and then the predictability of different feature groups was examined. Finally, we conducted an interpretive analysis to identify relatively important predictors.
Findings
The experimental results show that the random classifier with all extracted features outperformed other built models and achieved excellent performance, indicating the usefulness of these factors in predicting the donations. Moreover, the predictability of content features was demonstrated to be relatively better than that of creator ones. Finally, several particularly important predictors were identified such as the number of modal particles in the article.
Originality/value
This study is among the first to investigate what factors might drive customers' voluntary donation to content contributors on social websites. Different from previous studies focusing on live video streaming, we expand the research vision by examining the donations to user-generated text content, calling for attention to other important topics in the burgeoning industry.
Details
Keywords
Peter Lugosi, Hania Janta and Pamela Watson
This paper aims to introduce the notion of investigative research on the internet (IRI) and conceptualise its processes through the principle of streaming. It seeks to discuss the…
Abstract
Purpose
This paper aims to introduce the notion of investigative research on the internet (IRI) and conceptualise its processes through the principle of streaming. It seeks to discuss the similarities and differences between IRI and netnography and considers various aspects of the IRI process, including site selection, sampling, data collection and analysis.
Design/methodology/approach
Investigative internet‐based research uses the techniques of ethnography and netnography, including variations of participant observation and analysis of visual and textual material. Three international empirical cases are used to illustrate the application of IRI and streaming in research on international workers, consumer cultures and on emerging business phenomena.
Findings
IRI has a number of potential applications for hospitality management academics and practitioners. Streaming can help to understand the processes involved in conducting netnographic research, and streaming is a more appropriate way to conceptualise some internet‐based studies that do not conform to netnographic or ethnographic ideals.
Research limitations/implications
The three empirical cases highlight the processes of streaming in practice, which can be applied elsewhere. Principal limitations are the ethical dimensions of conducting undisclosed research and the sampling bias resulting from adopting an unobtrusive role and focusing on active internet users.
Practical implications
The paper highlights several issues, identified through streaming, that can be used to design human resource, marketing and operational strategies.
Originality/value
The paper demonstrates the application of streaming. Streaming can help researchers conduct netnographic studies; it is also a more appropriate way to describe broader types of investigative internet research. Moreover, it demonstrates the applicability of streaming in research on hospitality management and public policy issues.
Details
Keywords
Mary K. VanUllen, Emily Mock and Emmalyn Rogers
The purpose of this study is to examine the options for streaming video service available to libraries and determine which platform would best fit the needs of the University at…
Abstract
Purpose
The purpose of this study is to examine the options for streaming video service available to libraries and determine which platform would best fit the needs of the University at Albany Libraries.
Design/methodology/approach
Usage data and faculty and student feedback about the streaming video collections already in use by the libraries were compiled to evaluate current needs, and information was gathered about a selection of additional streaming video platforms to be considered.
Findings
It was determined that a multi-disciplinary collection with a patron-driven-style subscription model would be the best choice to add to the libraries streaming video offerings.
Research limitations/implications
This study focuses on the needs and experiences of the University at Albany Libraries, but the methodology can be used by other institutions assessing their own collections.
Originality/value
Most of the current literature related to streaming video in libraries focuses on building new collections, with little discussion of adding to existing collections – a gap which this study aims to fill.
Details
Keywords
Shih-Wei Chou, Ming-Chia Hsieh and Hui-Chun Pan
This study aims to understand how information-sharing in live-streaming is formed through a motivational perspective. The authors provide a framework to explain how live-streaming…
Abstract
Purpose
This study aims to understand how information-sharing in live-streaming is formed through a motivational perspective. The authors provide a framework to explain how live-streaming services and attachment affect viewers' information-sharing decision.
Design/methodology/approach
This study uses a survey-based method to collect data and partial least squares to analyze them.
Findings
The proposed hypotheses are largely supported. The results show that information-sharing intention is influenced by both attachment to a creator and attachment to a group. These attachments are positively affected by live-streaming services. The findings contribute to live-streaming literature by conceptualizing motivation and motivational feedback as service and attachment respectively.
Practical implications
The findings suggest that live-streaming managers emphasize social-technical features and relationship development with others (creators, group members) to motivate viewers' participation in live-streaming.
Originality/value
This study addresses the gap of lacking a systematic consideration of motivation in the live-streaming context. As such, the authors conducted empirical research that describes the information-sharing through the motivation from service and feedback from attachment.
Details
Keywords
Johann N. Giertz, Welf H. Weiger, Maria Törhönen and Juho Hamari
Social live-streaming services are an emerging form of social media that is gaining in popularity among researchers and practitioners. By facilitating real-time interactions…
Abstract
Purpose
Social live-streaming services are an emerging form of social media that is gaining in popularity among researchers and practitioners. By facilitating real-time interactions between video content creators (i.e. streamers) and viewers, live-streaming platforms provide an environment for novel engagement behaviors and monetization structures. This research aims to examine communication foci and styles as levers of streaming success. In doing so, the authors analyze their impact on viewers' engagement with the stream.
Design/methodology/approach
This research draws on a unique dataset collected via a multi-wave questionnaire comprising viewers' perceptions of a specific streamer's communications and their actual behavior toward them. The authors analyze the proposed impact of communication foci on viewing and donating behavior while considering the moderating role of communication style using seemingly unrelated regressions.
Findings
The results show that communication foci represent a double-edged sword: community-focused communication drives viewership while reducing donations made to the streamer. By contrast, content-focused communication curbs viewing but drives donating.
Practical implications
Of specific interest for practitioners, the study demonstrates how streaming content providers (e.g. influencers) should adjust their communications to drive engagement in the context of synchronous social media such as social live-streaming services. Beyond that, this research identifies unique characteristics of engagement that can help managers to improve their digital service offerings.
Social implications
Social live-streaming services provide an environment that offers unique opportunities for self-development and co-creation among social media users. By allowing for real-time interactions, these emerging social media services build on ephemeral content to provide altered experiences for users.
Originality/value
The authors highlight the need to distinguish between engagement behaviors in asynchronous and synchronous social media. The proposed conceptualization sheds new light on success factors of social media in general and social live-streaming services specifically. To maximize user engagement, content creators in synchronous social media must consider their communications' focus (content or community) and style (utilitarian or hedonic).
Details
Keywords
Anita K. Foster and Gene R. Springs
Academic libraries are struggling to support the growing demand for streaming video. The purpose of this paper is to detail the experience of running three long-term pilots with…
Abstract
Purpose
Academic libraries are struggling to support the growing demand for streaming video. The purpose of this paper is to detail the experience of running three long-term pilots with different streaming video platforms, including processes involved, lessons learned and next steps.
Design/methodology/approach
This paper uses a mixed methods approach, combining analysis of usage data with case study observations.
Findings
The length of the pilots allowed for deep understanding of the needs of this academic library’s community’s engagement with streaming video in the classroom, and confirmed anecdotal information that availability of multiple platforms supports diverse needs which led to continuing access to all platforms, operationalized to be managed within existing processes. Using usage data and feedback from a task force led to decisions to continue with all three platforms that were piloted.
Research limitations/implications
While this research describes the experience at one academic library, the information may be generalizable enough that other libraries may use it for their streaming video collection development decisions.
Originality/value
Long-term pilot studies for streaming video platforms can be challenging for many libraries to undertake. With a modest initial financial commitment, the library was able to explore how the community might use streaming video. Through analysis of usage data, the library was able to see when, where and what was being used and could make better informed decisions about where to concentrate future funds for streaming video support.
Details
Keywords
Daniel Yi Xiao, Barbara A. Pietraszewski and Susan P. Goodwin
As the use of electronic library resources increases, the demand for online support also multiplies. Information literacy and 24/7 customer support are some of the urgent issues…
Abstract
As the use of electronic library resources increases, the demand for online support also multiplies. Information literacy and 24/7 customer support are some of the urgent issues related to research in an electronic environment that many libraries are trying to address today. This article describes an approach in meeting these challenges, the Let‐It‐V (Learning E‐Resources Through Instructional Technology Videos) project at the Texas A&M University Libraries. This study combines the use of screen‐captured videos and a streaming media encoder to produce topic‐specific videos for task‐oriented demands. It is visual, interactive, and seeks to provide just‐in‐time solutions at a point of need. On‐demand streaming is a viable, cost‐effective alternative for low bandwidth delivery of video‐enabled library instruction. The technologies involved, key development issues, lessons learned and their implications for distance learning are discussed.
Details
Keywords
Tulsi Pawan Fowdur, M.A.N. Shaikh Abdoolla and Lokeshwar Doobur
The purpose of this paper is to perform a comparative analysis of the delay associated in running two real-time machine learning-based applications, namely, a video quality…
Abstract
Purpose
The purpose of this paper is to perform a comparative analysis of the delay associated in running two real-time machine learning-based applications, namely, a video quality assessment (VQA) and a phishing detection application by using the edge, fog and cloud computing paradigms.
Design/methodology/approach
The VQA algorithm was developed using Android Studio and run on a mobile phone for the edge paradigm. For the fog paradigm, it was hosted on a Java server and for the cloud paradigm on the IBM and Firebase clouds. The phishing detection algorithm was embedded into a browser extension for the edge paradigm. For the fog paradigm, it was hosted on a Node.js server and for the cloud paradigm on Firebase.
Findings
For the VQA algorithm, the edge paradigm had the highest response time while the cloud paradigm had the lowest, as the algorithm was computationally intensive. For the phishing detection algorithm, the edge paradigm had the lowest response time, and the cloud paradigm had the highest, as the algorithm had a low computational complexity. Since the determining factor for the response time was the latency, the edge paradigm provided the smallest delay as all processing were local.
Research limitations/implications
The main limitation of this work is that the experiments were performed on a small scale due to time and budget constraints.
Originality/value
A detailed analysis with real applications has been provided to show how the complexity of an application can determine the best computing paradigm on which it can be deployed.