No abstract

Unsolicited email campaigns remain as one of the biggest threats affecting millions of users per day. During the past years several techniques to detect unsolicited emails have been developed. This work provides means to validate the hypothesis that the identification of the email messages’ intention can be approached by sentiment analysis and personality recognition techniques. These techniques will provide new features that improve current spam classification techniques. We combine personality recognition and sentiment analysis techniques to analyse email content. We enrich a publicly available dataset adding these features, first separately and after in combination, of each message to the dataset, creating new datasets. We apply several combinations of the best email spam classifiers and filters to each dataset in order to compare results.

In the same manner that Online Social Networks (OSN) usage increases, non-legitimate campaigns over these types of web services are growing. This is the reason why significant number of users are affected by social spam every day and therefore, their privacy is threatened. To deal with this issue in this study we focus on mood analysis, among all content-based analysis techniques. We demonstrate that using this technique social spam filtering results are improved. First, the best spam filtering classifiers are identified using a labeled dataset consisting of Youtube comments, including spam. Then, a new dataset is created adding the mood feature to each comment, and the best classifiers are applied to it. A comparison between obtained results with and without mood information shows that this feature can help to improve social spam filtering results: the best accuracy is improved in two different datasets, and the number of false positives is reduced 13.76% and 11.41% on average. Moreover, the results are validated carrying out the same experiment but using a different dataset.

Despite new developments in machine learning classification techniques, improving the accuracy of spam filtering is a difficult task due to linguistic phenomena that limit its effectiveness. In particular, we highlight polysemy, synonymy, the usage of hypernyms/hyponyms, and the presence of irrelevant/confusing words. These problems should be solved at the pre-processing stage to avoid using inconsistent information in the building of classification models. Previous studies have suggested that the use of synset-based representation strategies could be successfully used to solve synonymy and polysemy problems. Complementarily, it is possible to take advantage of hyponymy/hypernymy-based to implement dimensionality reduction strategies. These strategies could unify textual terms to model the intentions of the document without losing any information (e.g., bringing together the synsets “viagra”, “ciallis”, “levitra” and other representing similar drugs by using “virility drug” which is a hyponym for all of them). These feature reduction schemes are known as lossless strategies as the information is not removed but only generalised. However, in some types of text classification problems (such as spam filtering) it may not be worthwhile to keep all the information and let dimensionality reduction algorithms discard information that may be irrelevant or confusing. In this work, we are introducing the feature reduction as a multi-objective optimisation problem to be solved using a Multi-Objective Evolutionary Algorithm (MOEA). Our algorithm allows, with minor modifications, to implement lossless (using only semantic-based synset grouping), low-loss (discarding irrelevant information and using semantic-based synset grouping) or lossy (discarding only irrelevant information) strategies. The contribution of this study is two-fold: (i) to introduce different dimensionality reduction methods (lossless, low-loss and lossy) as an optimization problem that can be solved using MOEA and (ii) to provide an experimental comparison of lossless and low-loss schemes for text representation. The results obtained support the usefulness of the low-loss method to improve the efficiency of classifiers.