With the exponential increase of social media users, cyberbullying has been emerged as a form of bullying through electronic messages. Cyberbullying detection is generally in social networks like Twitter is one of the focussed research area. Cyberbullying is serious and widespread issues affecting increasingly more Internet users. Text mining tools are detecting cyber bullying and deal with several issues. However the existing system has issue with time consumption and inaccurate Cyberbullying detection results for the given Twitter dataset. To avoid the above mentioned issues, in this work, Enhanced Cuckoo Search optimization (ECSO) and Hybrid Firefly Artificial Neural Network (HFANN) algorithm is proposed. The proposed system contains three main phases are such as preprocessing, feature subset selection and classification. The preprocessing is done by using k-means algorithm for reducing the noise data from the given Twitter dataset. It handles the missing features and redundancy features through k-means centroid values and min max normalization respectively. It is used to increase the classification accuracy more effectively. The pre-processed features are taken into feature selection process for obtaining more informative features from the Twitter dataset. It is performed by using ECSO algorithm and the objective function is used to compute the relevant and important feature based on the best fitness values. Then the HFANN algorithm is applied for classification through training and testing model. It classifies the features more accurately using best fireflies rather than the previous algorithms. The experimental result proves that the proposed ECSO+HFANN algorithm provides better classification performance in terms of lower time complexity, higher precision, recall, f-measure and accuracy than the existing algorithms.
Using a Feature Subset Selection method and Support Vector Machine to address curse of dimensionality and redundancy in Hyperion hyperspectral data classification
