The Internet is an enabling technology that assists daily and business activities. However, a network fault could prevent the user from accessing the internet thus creating trouble tickets. Ideally, accurate prediction prior to network fault allows the telco to respond before the customer raises a trouble ticket. Current research focuses on forecasting the quantity of trouble ticket using historical trouble ticket. To improve the prediction of network fault, the customer trouble ticket data is augmented to include internet usage data and signal measurement data. Random Forest (RF) and C5.0 Decision Tree algorithms are used to derive predictive models. Experiment results reveal that RF shows higher AUC score as compared to C5.0 Decision Tree. RF is able to identify the important features while C5.0 Decision Tree is able to list decision rules that describe the relation among selected features.

Nowadays, with the use of technology and the Internet, it is easy to start a business, more specifically an e-commerce business. However, maintaining a consistent sale and having returning customers can prove a challenge as most businesses rely on new customers for profits and does not generate a reliable profit as compared to relying on old customers. One might resort to applying different kinds of marketing strategies but without understanding of their customer base and proper segmentation of customers, these efforts could result in waste of resources and low probability of success. Therefore, an approach named J-WS that can perform customer segmentation based on customer sales data and Recency, Frequency, and Monetary (RFM) model is proposed. Meaningful information from different groups of customers can later be utilized by target marketing strategy to improve customer retention and impactful marketing. The proposed work consists of 5 phases which include data cleaning, identifying the best clustering algorithm between K-Means and Hierarchical clustering in terms of execution time and Sum of Squared Error, applying association rule mining to generate sets of frequent association rules among the clusters. Conclusively, J-WS can be used by e-commerce businesses to segment their customers meaningfully and properly.

Due to the detrimental consequences caused by cyberbullying, a great deal of research has been undertaken to propose effective techniques to resolve this reoccurring problem. The research presented in this paper is motivated by the fact that negative emotions can be caused by cyberbullying. This paper proposes cyberbullying detection models that are trained based on contextual, emotions and sentiment features. An Emotion Detection Model (EDM) was constructed using Twitter datasets that have been improved in terms of its annotations. Emotions and sentiment were extracted from cyberbullying datasets using EDM and lexicons based. Two cyberbullying datasets from Wikipedia and Twitter respectively were further improved by comprehensive annotation of emotion and sentiment features. The results show that anger, fear and guilt were the major emotions associated with cyberbullying. Subsequently, the extracted emotions were used as features in addition to contextual and sentiment features to train models for cyberbullying detection. The results demonstrate that using emotion features and sentiment has improved the performance of detecting cyberbullying by 0.5 to 0.6 recall. The proposed models also outperformed the state-of-the-art models by a 0.7 f1-score. The main contribution of this work is two-fold, which includes a comprehensive emotionannotated dataset for cyberbullying detection, and an empirical proof of emotions as effective features for cyberbullying detection.INDEX TERMS Cyberbullying, BERT, emotion mining, sentiment analysis.