Exploring high-level features for detecting cyberpedophilia

Bogdanova, Dasha; Rosso, Paolo; Solorio, Thamar

doi:10.1016/j.csl.2013.04.007

Cited by 46 publications

(41 citation statements)

References 12 publications

(20 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…These type of conversations pose serious challenges to systems which only focus on the predator-victim characterisation since in such systems the majority of features involves sexual content. The use of stages for the characterisation of predator conversations could potentially help systems in reducing the number of false positives when exposed to non-peadophile conversations with sexual content since predators' luring stages are not common in standard online sexual conversations [2].…”

Section: Discussionmentioning

confidence: 99%

“…To compute the sentiment polarity of a sentence we use Sentistrength. 2 The features used to characterise content, psycho-linguistic and discourse patterns are a bit more complex and will therefore be explained in more detail in the following subsections.…”

Section: Semantic Framesmentioning

confidence: 99%

“…In [2], Bogdanova et al, approach the problem of discriminating cyber-sex conversations from child grooming conversations by characterising them using n-grams and high-level features. Such features include emotion, neurotism, and those proposed by Michalopoulos et al [17].…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

Detecting Child Grooming Behaviour Patterns on Social Media

Cano

Fernández

Alani

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Online paedophile activity in social media has become a major concern in society as Internet access is easily available to a broader younger population. One common form of online child exploitation is child grooming, where adults and minors exchange sexual text and media via social media platforms. Such behaviour involves a number of stages performed by a predator (adult) with the final goal of approaching a victim (minor) in person. This paper presents a study of such online grooming stages from a machine learning perspective. We propose to characterise such stages by a series of features covering sentiment polarity, content, and psycho-linguistic and discourse patterns. Our experiments with online chatroom conversations show good results in automatically classifying chatlines into various grooming stages. Such a deeper understanding and tracking of predatory behaviour is vital for building robust systems for detecting grooming conversations and potential predators on social media.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Semantic Framesmentioning

confidence: 99%

See 1 more Smart Citation

Detecting Child Grooming Behaviour Patterns on Social Media

Cano

Fernández

Alani

2014

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…Other features, such as whether a profile image represents the user truthfully (Hancock and Toma, 2009), similarity of attributes such as name, given the Levenshtein difference (Li and Wang, 2015), and the emotional state of a user given the content they post (Bogdanova et al, 2014), have also been posited as being potentially useful for detecting identity deception.…”

Section: Similar Identity Attributes As Those Depicted Inmentioning

confidence: 99%

Cyber-security: Identity deception detection on social media platforms

Walt

Eloff

Grobler

2018

Computers & Security

View full text Add to dashboard Cite

Social media platforms allow billions of individuals to share their thoughts, likes and dislikes in real-time, without any censorship. This freedom, however, comes at a cyber-security risk. Cyber threats are more difficult to detect in a cyber world where anonymity and false identities are ever-present. The speed at which these deceptive identities evolve calls for solutions to detect identity deception. Cybersecurity threats caused by humans on social media platforms are widespread and warrant attention. This research posits a solution towards the intelligent detection of deceptive identities contrived by human individuals on social media platforms (SMPs). Firstly, this research evaluates machine learning models by using attributes such as the "profile image" found on SMPs. To improve on the results delivered by these models, past research findings from the field of psychology, such as that humans lie about their gender, are used. Newly engineered features such as "gender-derived-from-the-profileimage" are evaluated to grasp whether these features detect deception with greater accuracy. Furthermore, research results from detecting non-human (also known as bot) accounts are also leveraged to improve on the initial results. These machine learning results are lastly applied to a proposed model for the intelligent detection and interpretation of identity deception on SMPs. This paper shows that the cyber-security threat of identity deception can potentially be minimized, should the vulnerability in the current way of setting up user accounts on SMPs be re-engineered in the future.

show abstract

“…Profiling anonymous authors is a problem of growing importance, both from forensic and marketing perspectives. From a forensic perspective it is important to identify the linguistic profile of an author of a harassing text message or a potential online paedophile on the basis of the analysis of his writing style in order, for instance, to unveil his age [58] [7]. From a marketing viewpoint, companies may be interested in knowing the demographics of their target group in order to achieve a better market segmentation.…”

Section: Author Profiling: How Writing Style Is Sharedmentioning

confidence: 99%