Kevin Chai

Content Quality Assessment Related Frameworks for Social Media

The assessment of content quality (CQ) in social media adds a layer of complexity over traditional information quality assessment frameworks. Challenges arise in accurately evaluating the quality of content that has been created by users from different backgrounds, for different domains and consumed by users with different requirements. This paper presents a comprehensive review of 19 existing CQ assessment related frameworks for social media in addition to proposing directions for framework improvements.

Web Spambot Detection Based on Web Navigation Behaviour

Web robots have been widely used for various beneficial and malicious activities. Web spambots are a type of web robot that spreads spam content throughout the web by typically targeting Web 2.0 applications. They are intelligently designed to replicate human behaviour in order to bypass system checks. Spam content not only wastes valuable resources but can also mislead users to unsolicited websites and award undeserved search engine rankings to spammers’ campaign websites. While most of the research in anti-spam filtering focuses on the identification of spam content on the web, only a few have investigated the origin of spam content, hence identification and detection of web spambots still remains an open area of research. In this paper, we describe an automated supervised machine learning solution which utilises web navigation behaviour to detect web spambots. We propose a new feature set (referred to as an action set) as a representation of user behaviour to differentiate web spambots from human users. Our experimental results show that our solution achieves a 96.24% accuracy in classifying web spambots.

HoneySpam 2.0: Profiling Web Spambot Behaviour

Internet bots have been widely used for various beneficial and malicious activities on the web. In this paper we provide new insights into a new kind of bot termed as web spambot which is primarily used for spreading spam content on the web. To gain insights into web spambots, we developed a tool (HoneySpam 2.0) to track their behaviour. This paper presents two main contributions, firstly it describes the design of HoneySpam 2.0 and secondly we outline the experimental results that characterise web spambot behaviour. By profiling web spambots, we provide the foundation for identifying such bots and preventing and filtering web spam content.

A survey of revenue models for current generation social software's systems

In this paper we survey a number of different social software websites and analyze their revenue models. The main revenue models that we analyzed included - advertising, premium memberships, affiliate programs, donations and merchandize sale. The survey will categorize different social software based upon the revenue model being adopted. The main aim of this paper is to highlight the need for studying revenue sharing models which attempt to reward the users in the online community for participating in a particular social software website. A total of 7 revenue sharing websites would be discussed to show the importance of revenue sharing.

Behaviour-Based web spambot detection by utilising action time and action frequency

Web spam is an escalating problem that wastes valuable resources, misleads people and can manipulate search engines in achieving undeserved search rankings to promote spam content. Spammers have extensively used Web robots to distribute spam content within Web 2.0 platforms. We referred to these web robots as spambots that are capable of performing human tasks such as registering user accounts as well as browsing and posting content. Conventional content-based and link-based techniques are not effective in detecting and preventing web spambots as their focus is on spam content identification rather than spambot detection. We extend our previous research by proposing two action-based features sets known as action time and action frequency for spambot detection. We evaluate our new framework against a real dataset containing spambots and human users and achieve an average classification accuracy of 94.70%.

Assessing post usage for measuring the quality of forum posts

It has become difficult to discover quality content within forums websites due to the increasing amount of User Generated Content (UGC) on the Web. Many existing websites have relied on their users to explicitly rate content quality. The main problem with this approach is that the majority of content often receives insufficient rating. Current automated content rating solutions have evaluated linguistic features of UGC but are less effective for different types of online communities. We propose a novel approach that assesses post usage to measure the quality of forum posts. Post usage can be viewed as implicit user ratings derived from their usage behaviour. The proposed model is validated against an operational forum using Matthews Correlation Coefficient to measure performance. Our model serves as a basis of exploring content usage to measure content quality in forums and other Web 2.0 platforms.

User contribution measurement model for web-based discussion forums

The success of social software depends on contributions made by two key entities; the infrastructure provider(s) and the content providers (users). Currently, social software providers do not possess a powerful and generic approach to measure the contributions of their users. The ability of measuring user contributions will allow social software providers to accurately identify, acknowledge and reward their content contributors. As a result, content providers may become motivated to contribute content more regularly. This paper proposes a user contribution measurement model which is validated against an operational web-based discussion forum.

How much money do spammers make from your website

Despite years of researchers contribution in the domain of spam filtering, the question as to how much money spammers can make has largely remained unanswered. The value of spam-marketing on the web can be determined by discovering the cost of distributing spam in Web 2.0 platforms, and the success ratio of turning a spamming campaign into a profitable activity. Currently, there is limited knowledge on the nature of spam distribution in web applications, and public methods for estimating the turnover rate for spammers, in the existing literature. Therefore, we adopted a methodological approach to address these issues and measure the value of spam-marketing on the web. Using current spam tactics, we targeted 66,226 websites both in English and non-English languages. We launched a spam campaign and set up a website to replicate spam practices. We posted spam content to 7,772 websites that resulted in 2059 unique visits to our website, and 3 purchase transactions, in a period of a month. The total conversion visit rate for this experiment was 26.49% and purchase rate was 0.14%.

Image and text relations in ISIS materials and the new relations established through recontextualisation in online media

This study takes a systemic functional multimodal social semiotic approach to the analysis and discussion of image and text relations in two sets of data. First, patterns of contextualisation of images and text in the online magazines Dabiq and Rumiyah produced by the Islamic extremist organisation which refers to itself as Islamic State (referred to here as ISIS) are examined. The second data set consists of a sample of texts from Western online news and blog sites which include recontextualisations of images found in the first data set. A sample of examples of the use and re-use of images is discussed in order to identify patterns of similarity and difference when images and text are recontextualised. It is argued that the ISIS material tends to foreground the interpersonal metafunction in combination with the textual metafunction (i.e. the stance towards the content and the organisation of the message for this purpose), while the other data set tends to foreground the ideational metafunction (the participants, processes and circumstances of what is being reported). These inferences indicate that further exploration of a larger data set is worth pursuing. Such studies would provide deeper insights helping to distinguish between online material which supports terrorism and that which opposes it, as well as facilitating the further development of multimodal social semiotic approaches to image and text relations.