Spam Posts Research Articles

On unlinkability and denial of service attacks resilience of whistleblower platforms

This work explores how to enhance pseudonymous whistleblower submission systems, specifically by supporting protocol level unlinkability, while also making the system resilient against (distributed) denial of service attacks. To that end, we propose a blind signature based protocol which facilitates assignment of trust to anonymous posters in a manner which depends on the quality of prior posts, yet unlinkable to said posts or corresponding poster. This (multi-level) trust is leveraged to prioritize the posts, thus mitigating the effect that spam posts may have on the party reviewing the posts. We design and carry out simulations to explore the resilience of the whistleblower submission system against denial of service attacks while applying the proposed approach. Our experiments affirm that for a range of realistic scenarios the proposed approach provides reasonable mitigation.

A Content Filtering from Spam Posts on Social Media using Weighted Multimodal Approach

The system for filtering spam posts on social media is preferred to obtain the relevant content and expected by users. The previous works on spam detection have been done to filter irrelevant content on email and social media based on text or image separately. Due to the social media posts are commonly in the form of image, text, or both, the multimodal data is preferred to improve the capability of system in handling filtering content on social media. In addition, a spam post containing multimodal data sometimes does not indicate spam in both data but only one. To improve the performance of system, we propose a weighted multimodal approach for filtering content from spam posts in social media using Convolutional Neural Network (CNN). The mechanism of weighted multimodal is by weighting of spam prediction results from image and text data. We also investigate the performance of CNN architectures for spam post detection that are 3-layer, 5-layer, AlexNet and VGG16. The performance of each architectures is evaluated by 8000 Indonesian posts in the form of image and text taken from Instagram posts. The results show that the highest accuracy achieves 0.9850 based on the combination of image and text by using a 5-layer architecture. The average accuracy of all CNN architectures using multimodal data is higher than only using image and text data separately.

Content Noise Detection Model Using Deep Learning in Web Forums

Spam posts in web forum discussions cause user inconvenience and lower the value of the web forum as an open source of user opinion. In this regard, as the importance of a web post is evaluated in terms of the number of involved authors, noise distorts the analysis results by adding unnecessary data to the opinion analysis. Here, in this work, an automatic detection model for spam posts in web forums using both conventional machine learning and deep learning is proposed. To automatically differentiate between normal posts and spam, evaluators were asked to recognize spam posts in advance. To construct the machine learning-based model, text features from posted content using text mining techniques from the perspective of linguistics were extracted, and supervised learning was performed to distinguish content noise from normal posts. For the deep learning model, raw text including and excluding special characters was utilized. A comparison analysis on deep neural networks using the two different recurrent neural network (RNN) models of the simple RNN and long short-term memory (LSTM) network was also performed. Furthermore, the proposed model was applied to two web forums. The experimental results indicate that the deep learning model affords significant improvements over the accuracy of conventional machine learning associated with text features. The accuracy of the proposed model using LSTM reaches 98.56%, and the precision and recall of the noise class reach 99% and 99.53%, respectively.

Portraying Privacy Leakage of Public WiFi Systems for Users on Travel Spam Detection in Social Bookmarking System

In this paper, we depicts spam revelation, in perspective of the examination of posts, in social bookmarking districts. For consistent acknowledgment of spam posts, we propose a name estimation plot and a specific evaluation procedure for picking marks. The label estimation scores each tag. In the particular evaluation, the label scores in perspective of the utilization repeat and the degree of spammers are estimated and the thoughts of white tag and dim tag are introduced. Using these thoughts, names are proficiently arranged into the names demolishing the execution of spam revelation, the names pleasing in getting spammers, and the marks which should achieve a discipline. Finally, we propose semantic components to moreover upgrade the spam distinguishing proof

A Neural Network-Based Ensemble Approach for Spam Detection in Twitter

As the social networking sites get more popular, spammers target these sites to spread spam posts. Twitter is one of the most popular online social networking sites where users communicate and interact on various topics. Most of the current spam filtering methods in Twitter focus on detecting the spammers and blocking them. However, spammers can create a new account and start posting new spam tweets again. So there is a need for robust spam detection techniques to detect the spam at tweet level. These types of techniques can prevent the spam in real time. To detect the spam at tweet level, often features are defined, and appropriate machine learning algorithms are applied in the literature. Recently, deep learning methods are showing fruitful results on several natural language processing tasks. We want to use the potential benefits of these two types of methods for our problem. Toward this, we propose an ensemble approach for spam detection at tweet level. We develop various deep learning models based on convolutional neural networks (CNNs). Five CNNs and one feature-based model are used in the ensemble. Each CNN uses different word embeddings (Glove, Word2vec) to train the model. The feature-based model uses content-based, user-based, and n-gram features. Our approach combines both deep learning and traditional feature-based models using a multilayer neural network which acts as a meta-classifier. We evaluate our method on two data sets, one data set is balanced, and another one is imbalanced. The experimental results show that our proposed method outperforms the existing methods.

Attribute selection for improving spam classification in online social networks: a rough set theory-based approach

As online social network (OSN) sites become increasingly popular, they are targeted by spammers who post malicious content on the sites. Hence, it is important to filter out spam accounts and spam posts from OSNs. There exist several prior works on spam classification on OSNs, which utilize various features to distinguish between spam and legitimate entities. The objective of this study is to improve such spam classification, by developing an attribute selection methodology that helps to find a smaller subset of the attributes which leads to better classification. Specifically, we apply the concepts of rough set theory to develop the attribute selection algorithm. We perform experiments over five different spam classification datasets over diverse OSNs and compare the performance of the proposed methodology with that of several baseline methodologies for attribute selection. We find that, for most of the datasets, the proposed methodology selects an attribute subset that is smaller than what is selected by the baseline methodologies, yet achieves better classification performance compared to the other methods.

인스타그램 기반의 전이학습과 게시글 메타 정보를 활용한 페이스북 스팸 게시글 판별

This study develops a text spam filtering system for Facebook based on two variable categories: keywords learned from Instagram and meta-information of Facebook posts. Since there is no explicit labels for spam/ham posts, we utilize hash tags in Instagram to train classification models. In addition, the filtering accuracy is enhanced by considering meta-information of Facebook posts. To verify the proposed filtering system, we conduct an empirical experiment based on a total of 1,795,067 and 761,861 Facebook and Instagram documents, respectively. Employing random forest as a base classification algorithm, experimental result shows that the proposed filtering system yield 99% and 98% in terms of filtering accuracy and F1-measure, respectively. We expect that the proposed filtering scheme can be applied other web services suffering from massive spam posts but no explicit spam labels are available.

Modeling and Evaluating Information Diffusion for Spam Detection in Micro-blogging Networks

Spam has become one of the top threats of micro-blogging networks as the representations of rumor spreading, advertisement abusing and malware distribution. With the increasing popularity of micro-blogging, the problems will exacerbate. Prior detection tools are either designed for specific types of spams or not robust enough. Spammers may escape easily from being detected by adjusting their behaviors. In this paper, we present a novel model to quantitatively evaluate information diffusion in micro-blogging networks. Under this model, we found that spam posts differ wildly from the non-spam ones. First, the propagations of non-spam posts mostly result from their followers, but those of spam posts are mainly from strangers. Second, the non-spam posts relatively last longer than the spam posts. Besides, the non-spam posts always get their first reposts/comments much sooner than the spam posts. With the features defined in our model, we propose an RBF-based approach to detect spams. Different from the previous works, in which the features are extracted from individual profiles or contents, the diffusion features are not determined by any single user but the crowd. Thus, our method is more robust because any single user’s behavior changes will not affect the effectiveness. Besides, although the spams vary in types and forms, they’re propagated in the same way, so our method is effective for all types of spams. With the real data crawled from the leading micro-blogging services of China, we are able to evaluate the effectiveness of our model. The experiment results show that our model can achieve high accuracy both in precision and recall.

상호작용성에 의한 SNS 영향유저 선정에 관한 연구 : 연속적인 참조관계가 있는 블로고스피어를 중심으로

Various influence-related relationships in Social Network Services (SNS) among users, posts, and user-and-post, can be expressed using links. The current research evaluates the influence of specific users or posts by analyzing the link structure of relevant social network graphs to identify influential users. We applied the concept of mutual interactions proposed for ranking semantic web resources, rather than the voting notion of Page Rank or HITS, to blogosphere, one of the early SNS. Through many experiments with network models, where the performance and validity of each alternative approach can be analyzed, we showed the applicability and strengths of our approach. The weight tuning processes for the links of these network models enabled us to control the experiment errors form the link weight differences and compare the implementation easiness of alternatives. An additional example of how to enter the content scores of commercial or spam posts into the graph-based method is suggested on a small network model as well. This research, as a starting point of the study on identifying influential users in SNS, is distinctive from the previous researches in the following points. First, various influence-related properties that are deemed important but are disregarded, such as scraping, commenting, subscribing to RSS feeds, and trusting friends, can be considered simultaneously. Second, the framework reflects the general phenomenon where objects interacting with more influential objects increase their influence. Third, regarding the extent to which a bloggers causes other bloggers to act after him or her as the most important factor of influence, we treated sequential referencing relationships with a viewpoint from that of PageRank or HITS (Hypertext Induced Topic Selection).

Malicious and Spam Posts in Online Social Networks

A large-scale study of more than half a million Facebook posts suggests that members of online social networks can expect a significant chance of encountering spam posts and a much lower but not negligible chance of coming across malicious links.

The Impact of Blog Recommendations on Security Prices and Trading Volumes

Previous research and recent statements by the Securities and Exchange Commission indicate that e-mail spam and message board posts are often used to manipulate markets in a variation of the classical pump-and-dump scheme, leading to temporary market reactions followed by price reversals. In contrast, I hypothesize that financial blogs spread genuine information leading to permanent market adjustments. I investigate stock recommendations on blogs and find that bloggers tend to write about liquid securities issued by large firms; I also find that bloggers offer short advice consistent with momentum strategies but long advice consistent with contrarian strategies. To test the hypothesis that stock recommendations on blogs impact both prices and trading volumes of the touted securities, I collect recommendations from blogs and analyze returns and trading volumes on the days surrounding publication; I offer evidence to support my hypothesis and I find some evidence of market reaction being stronger for short recommendations. I hypothesize and test that the magnitude of the market reaction depends on the market capitalization of the touted firm, on the size of the blog's audience, on the depth of the analysis of the blog post and on the perceived skill of the blog's author. I find that the market appears to react more strongly to recommendations given by holders of a graduate degree in Finance or Economics; the other hypothesized factors do not have an impact on the magnitude of the reaction. Finally, I document the absence of price reversals in the twenty days following blog publication, giving support to the hypothesis that blogs offer genuine information.