Conclusion and Future Directions

In this chapter, the authors summarize the research performed so far on automatic cyberbullying detection, which was the topic of this book. They summarize all chapters of the book. Next, they propose a general discussion of the potential and limitations of technology-based methods for detecting and preventing cyberbullying. They also ask what the ethical boundaries are for such technology to operate in everyday life. Should we allow constant surveillance for the sake of safety? Should we limit the technology, thus taking away its true problem-solving potential to match our freedom of speech? Or is there a third way in which both the technology is used to meet its potential, while not taking away the feeling of freedom? Let questions which arise during this last chapter become guidelines for future research on automatic detection and mitigation of cyberbullying.

Cyberbullying Blocker Test Application for Android Devices

In this chapter, the authors present an application for Android smartphones to automatically detect possible harmful content in input text. The developed application is aimed to test in practice the performance of the developed cyberbullying detection methods described in previous chapters. The final goal of the developed application will be to help mitigate the problem of cyberbullying by quickly detecting possibly harmful contents in user's entry and warning the user of the possible negative influence. The test application was prepared to use one of two methods for detection of harmful messages: a method inspired by a brute force search algorithm applied to language modelling and a method which uses seed words from three categories to calculate semantic orientation score SO-PMI-IR and then maximize the relevance of categories to specify harmfulness of a message (both methods were described in previous chapters). First tests showed that both methods are working properly under the Android environment.

Introduction

In this chapter, the authors present an introduction to this book. They explain the motivation that drove them in the research and draft an outline of the whole book.

Traditional Classifiers vs. Deep Learning for Cyberbullying Detection

In this chapter, the authors present their approach to cyberbullying detection with the use of various traditional classifiers, including a deep learning approach. Research has tackled the problem of cyberbullying detection during recent years. However, due to complexity of language used in cyberbullying, the results obtained with traditional classifiers has remained only mildly satisfying. In this chapter, the authors apply a number of traditional classifiers, used also in previous research, to obtain an objective view on to what extent each of them is suitable to the task. They also propose a novel method to automatic cyberbullying detection based on convolutional neural networks and increased feature density. The experiments performed on actual cyberbullying data showed a major advantage of the presented approach to all previous methods, including the two best performing methods so far based on SO-PMI-IR and brute-force search algorithm, presented in previous two chapters.

Description and Initial Analysis of Cyberbullying Dataset

In this chapter, the authors focus on datasets used in cyberbullying detection research. They describe and compare several datasets applied in previous research and describe in detail the dataset that they decided to apply in their research. They also perform an initial analysis of the dataset to find various characteristics. They preprocess the dataset in several ways for further use and perform affect analysis to find out whether emotion-related features tend to be characteristic for cyberbullying. Based on the results of affect analysis, they also perform an initial attempt to classify cyberbullying data using a simple machine learning approach, which will be considered as a baseline in forthcoming chapters.

Background

In this chapter, the authors present the background for the research described in this book. They describe the problem of cyberbullying in general, with specific reference to its status quo in Japan. They also describe some of the most relevant previous research done on the topic of cyberbullying detection and point out research gaps that they aim to fill with this book.

Brute Force Search Method for Cyberbullying Detection

In this chapter, the authors present a method for automatic detection of malicious internet contents, based on a combinatorial approach resembling brute force search algorithms, with application to language classification. The method automatically extracts sophisticated patterns from sentences and applies them in classification. The experiments performed on actual cyberbullying data showed advantage of this method to previous methods, including the one described in Chapter 4. Pros and cons of this method when compared to previous ones are also discussed in this chapter.

Web Mining-Based Method for Cyberbullying Detection

In this chapter, the authors present a method for automatic detection of cyberbullying entries based on a Web mining technique, in particular, on an extended SO-PMI-IR method calculating relevance of new input documents with training documents. The method uses seed words from three categories to calculate semantic orientation score and then maximizes the relevance of categories. The method outperformed previously proposed Web-mining-based methods in both laboratory and real-world conditions. The developed system is deployed and tested in practice. After a year of testing, the authors notice an over 30% point drop in its performance. They hypothesize on the reasons for the drop. To regain the lost performance and sustain it in the future, the authors propose additional improvements including automatic acquisition and filtering of seed words. Experimentally selected optimal improvements regained much of the lost performance.