Toxic comments on social media platforms, news portals, and online forums are impolite, insulting, or unreasonable that usually make other users leave a conversation. Due to the significant number of comments, it is impractical to moderate them manually. Therefore, online service providers use the automatic detection of toxicity using Machine Learning (ML) algorithms. However, the model's toxicity identification performance relies on the best combination of classifier and feature extraction techniques. In this empirical study, we set up a comparison environment for toxic comment classification using 15 frequently used supervised ML classifiers with the four most prominent feature extraction schemes. We considered the publicly available Jigsaw dataset on toxic comments written by human users. We tested, analyzed and compared with every pair of investigated classifiers and finally reported a conclusion. We used the accuracy and area under the ROC curve as the evaluation metrics. We revealed that Logistic Regression and AdaBoost are the best toxic comment classifiers. The average accuracy of Logistic Regression and AdaBoost is 0.895 and 0.893, respectively, where both achieved the same area under the ROC curve score (i.e., 0.828). Therefore, the primary takeaway of this study is that the Logistic Regression and Adaboost leveraging BoW, TF-IDF, or Hashing features can perform sufficiently for toxic comment classification.

Self-adaptive systems appeared in order to reduce the effort of manual software maintenance. Apart from software attributes, for example, different alternative software modules, self-adaptation decisions depend on environmental attributes, for example, service rate, bandwidth etc. Current well-known self-adaptation approaches can be further improved by incorporating environmental attributes. Moreover, reducing maintenance effort includes minimizing both operational and development effort. To reduce the effort of developing self-adaptive software, the constituent components should be reusable. This paper proposes a technique to incorporate environmental attributes to learning-based self-adaptation and to increase the reuse potential of self-adaptive system components. The environmental attributes are provided as a constraint to an optimization problem which results in an optimal software attribute selection. Design patterns for self-adaptive system components are proposed to improve its reusability. The proposed technique was validated on a news serving website called Znn.com. According to renowned reusability metrics such as Lines of Code (LOC), Message Passing Coupling (MPC) and Lack of Cohesion of Methods 4 (LCOM4), the proposed technique improved reuse potential. The website was further tested for adaptation effectiveness under two scenarios – adaptation and without adaptation. According to our experiments, Adaptation gradually improved the main goal response time of the website where it performed poorly without adaptation.