Cyberbullying refers to bullying and harassment of defenseless or vulnerable people such as children, teenagers, and women through any means of communication (e.g., e-mail, text messages, wall posts, tweets) over any online medium (e.g., social media, blogs, online games, virtual reality environments). The effect of cyberbullying may be severe and irreversible and it has become one of the major problems of cyber-societies in today’s electronic world. Prevention of cyberbullying activities as well as the development of timely response mechanisms require automated and accurate detection of cyberbullying acts. This study focuses on the problem of cyberbullying detection over Facebook activity content written in Turkish. Through extensive experiments with the various machine and deep learning algorithms, the best estimator for the task is chosen and then employed for both cross-domain evaluation and profiling of cyber-aggressive users. The results obtained with fivefold cross-validation are evaluated with an average-macro F1 score. These results show that BERT is the best estimator with an average macro F1 of 0.928, and employing it on various datasets collected from different OSN domains produces highly satisfying results. This article also reports detailed profiling of cyber-aggressive users by providing even more information than what is visible to the naked eye.

Privacy scoring aims at measuring the privacy violation risk of a user over an online social network (OSN) based on attribute values shared in the user’s OSN profile page and the user’s position in the network. Existing studies on privacy scoring rely on possibly biased or emotional survey data. In this study, we work with real-world data collected from the professional LinkedIn OSN and show that probabilistic scoring models derived from the item response theory fit real-world data better than naive approaches. We also introduce the granularity of the data an OSN user shares on her profile as a latent dimension of the OSN privacy scoring problem. Incorporating data granularity into our model, we build the most comprehensive solution to the OSN privacy scoring problem. Extensive experimental evaluation of various scoring models indicates the effectiveness of the proposed solution.

Abstract Sentiment analysis (SA) is an essential task for many domains where it is crucial to know users’ public opinion about events, products, brands, politicians and so on. Existing works on SA have concentrated on English texts including Twitter feeds and user reviews on hotels, movies and products. On the other hand, Facebook, as an online social network (OSN), has attracted quite limited attention from the research community. Among these, SA work on Turkish text obtained from OSNs are extremely scarce. In this paper, our aim is to perform SA on public Facebook data collected from Turkish user accounts. Our study differs from existing studies in terms of the data set scale, the natural language of the texts in the data set and the extent of experimental analyses that include both machine learning and deep learning techniques. We extensively report not only the results of different learning models involving SA but also statistical distribution of metadata of user activities across various user attributes (e.g. gender and age). Our experimental results indicate that recurrent neural networks achieve the best accuracy (i.e. 0.916) with word embeddings. To the best of our knowledge, this is the best result for SA on Facebook data in the context of the Turkish language.

Online Social Networks (OSNs) are very popular platforms for social interaction. Data posted publicly over OSNs pose various threats against the individual privacy of OSN users. Adversaries can try to predict private attribute values, such as gender, as well as links/connections. Quantifying an adversary’s capacity in inferring the gender of an OSN user is an important first step towards privacy protection. Numerous studies have been made on the problem of predicting the gender of an author/user, especially in the context of the English language. Conversely, studies in this field are quite limited for the Turkish language and specifically in the domain of OSNs. Previous studies for gender prediction of Turkish OSN users have mostly been performed by using the content of tweets and Facebook comments. In this article, we propose using various features, not just user comments, for the gender prediction problem over the Facebook OSN. Unlike existing studies, we exploited features extracted from profile, wall content, and network structure, as well as wall interactions of the user. Therefore, our study differs from the existing work in the broadness of the features considered, machine learning and deep learning methods applied, and the size of the OSN dataset used in the experimental evaluation. Our results indicate that basic profile information provides better results; moreover, using this information together with wall interactions improves prediction quality. We measured the best accuracy value as 0.982, which was obtained by combining profile data and wall interactions of Turkish OSN users. In the wall interactions model, we introduced 34 different features that provide better results than the existing content-based studies for Turkish.

Differential privacy (DP) has gained significant attention lately as the state of the art in privacy protection. It achieves privacy by adding noise to query answers. We study the problem of privately and accurately answering a set of statistical range queries in batch mode (i.e., under non-interactive DP). The noise magnitude in DP depends directly on the sensitivity of a query set, and calculating sensitivity was proven to be NP-hard. Therefore, efficiently bounding the sensitivity of a given query set is still an open research problem. In this work, we propose upper bounds on sensitivity that are tighter than those in previous work. We also propose a formulation to exactly calculate sensitivity for a set of COUNT queries. However, it is impractical to implement these bounds without sophisticated methods. We therefore introduce methods that build a graph model G based on a query set Q, such that implementing the aforementioned bounds can be achieved by solving two well-known clique problems on G. We make use of the literature in solving these clique problems to realize our bounds efficiently. Experimental results show that for query sets with a few hundred queries, it takes only a few seconds to obtain results.

Differential privacy has gained attention from the community as the mechanism for privacy protection. Significant effort has focused on its application to data analysis, where statistical queries are submitted in batch and answers to these queries are perturbed with noise. The magnitude of this noise depends on the privacy parameter ϵ and the sensitivity of the query set. However, computing the sensitivity is known to be NP-hard.

The integration of information dispersed among multiple repositories is a crucial step for accurate data analysis in various domains. In support of this goal, it is critical to devise procedures for identifying similar records across distinct data sources. At the same time, to adhere to privacy regulations and policies, such procedures should protect the confidentiality of the individuals to whom the information corresponds. Various private record linkage (PRL) protocols have been proposed to achieve this goal, involving secure multi-party computation (SMC) and similarity preserving data transformation techniques. SMC methods provide secure and accurate solutions to the PRL problem, but are prohibitively expensive in practice, mainly due to excessive computational requirements. Data transformation techniques offer more practical solutions, but incur the cost of information leakage and false matches. In this paper, we introduce a novel model for practical PRL, which 1) affords controlled and limited information leakage, 2) avoids false matches resulting from data transformation. Initially, we partition the data sources into blocks to eliminate comparisons for records that are unlikely to match. Then, to identify matches, we apply an efficient SMC technique between the candidate record pairs. To enable efficiency and privacy, our model leaks a controlled amount of obfuscated data prior to the secure computations. Applied obfuscation relies on differential privacy which provides strong privacy guarantees against adversaries with arbitrary background knowledge. In addition, we illustrate the practical nature of our approach through an empirical analysis with data derived from public voter records.

Real-world entities are not always represented by the same set of features in different data sets. Therefore, matching records of the same real-world entity distributed across these data sets is a challenging task. If the data sets contain private information, the problem becomes even more difficult. Existing solutions to this problem generally follow two approaches: sanitization techniques and cryptographic techniques. We propose a hybrid technique that combines these two approaches and enables users to trade off between privacy, accuracy, and cost. Our main contribution is the use of a blocking phase that operates over sanitized data to filter out in a privacy-preserving manner pairs of records that do not satisfy the matching condition. We also provide a formal definition of privacy and prove that the participants of our protocols learn nothing other than their share of the result and what can be inferred from their share of the result, their input and sanitized views of the input data sets (which are considered public information). Our method incurs considerably lower costs than cryptographic techniques and yields significantly more accurate matching results compared to sanitization techniques, even when privacy requirements are high.

Gaussian mixture models are an important tool in Bayesian decision theory. In this study, we focus on building such models over statistical database protected under differential privacy. Our approach involves querying necessary statistics from a database and building a Bayesian classifier over the noise added responses generated according to differential privacy. We formally analyze the sensitivity of our query set. Since there are multiple methods to query a statistic, either directly or indirectly, we analyze the sensitivities for different querying methods. Furthermore we establish theoretical bounds for the Bayes error for the univariate (one dimensional) case. We study the Bayes error for the multivariate (high dimensional) case in experiments with both simulated data and real life data. We discover that adding Laplace noise to a statistic under certain constraint is problematic. For example variance-covariance matrix is no longer positive definite after noise addition. We propose a heuristic method to fix the noise added variance-covariance matrix.

Private matching between datasets owned by distinct parties is a challenging problem with several applications. Private matching allows two parties to identify the records that are close to each other according to some distance functions, such that no additional information other than the join result is disclosed to any party. Private matching can be solved securely and accurately using secure multi-party computation (SMC) techniques, but such an approach is prohibitively expensive in practice. Previous work proposed the release of sanitized versions of the sensitive datasets which allows blocking, i.e., filtering out sub-sets of records that cannot be part of the join result. This way, SMC is applied only to a small fraction of record pairs, reducing the matching cost to acceptable levels. The blocking step is essential for the privacy, accuracy and efficiency of matching. However, the state-of-the-art focuses on sanitization based on k-anonymity, which does not provide sufficient privacy. We propose an alternative design centered on differential privacy, a novel paradigm that provides strong privacy guarantees. The realization of the new model presents difficult challenges, such as the evaluation of distance-based matching conditions with the help of only a statistical queries interface. Specialized versions of data indexing structures (e.g., kd-trees) also need to be devised, in order to comply with differential privacy. Experiments conducted on the real-world Census-income dataset show that, although our methods provide strong privacy, their effectiveness in reducing matching cost is not far from that of k-anonymity based counterparts.

In recent years, anonymization methods have emerged as an important tool to preserve individual privacy when releasing privacy sensitive data sets. This interest in anonymization techniques has resulted in a plethora of methods for anonymizing data under different privacy and utility assumptions. At the same time, there has been little research addressing how to effectively use the anonymized data for data mining in general and for distributed data mining in particular. In this paper, we propose a new approach for building classifiers using anonymized data by modeling anonymized data as uncertain data. In our method, we do not assume any probability distribution over the data. Instead, we propose collecting all necessary statistics during anonymization and releasing these together with the anonymized data. We show that releasing such statistics does not violate anonymity. Experiments spanning various alternatives both in local and distributed data mining settings reveal that our method performs better than heuristic approaches for handling anonymized data.

Real-world entities are not always represented by the same set of features in different data sets. Therefore matching and linking records corresponding to the same real-world entity distributed across these data sets is a challenging task. If the data sets contain private information, the problem becomes even harder due to privacy concerns. Existing solutions of this problem mostly follow two approaches: sanitization techniques and cryptographic techniques. The former achieves privacy by perturbing sensitive data at the expense of degrading matching accuracy. The later, on the other hand, attains both privacy and high accuracy under heavy communication and computation costs. In this paper, we propose a method that combines these two approaches and enables users to trade off between privacy, accuracy and cost. Experiments conducted on real data sets show that our method has significantly lower costs than cryptographic techniques and yields much more accurate matching results compared to sanitization techniques, even when the data sets are perturbed extensively.

Enterprises have been collecting data for many reasons including better customer relationship management, and high-level decision making. Public safety was another motivation for large-scale data collection efforts initiated by government agencies. However, such widespread data collection efforts coupled with powerful data analysis tools raised concerns about privacy. This is due to the fact that collected data may contain confidential information. One method to ensure privacy is to selectively hide confidential information from the data sets to be disclosed. In this paper, we focus on hiding confidential correlations. We introduce a heuristic to reduce the information loss and propose a blocking method that prevents discovery of confidential correlations while preserving the usefulness of the data set.