This paper aims to present the design and implementation of a prototype that recognizes grooming attacks in the context of COP (child online protection) using Natural Language Processing and Machine Learning hybrid model, via Convolutional Neural Networks (CNN). The solution uses a vector representation of words as the semantic model and the implementation of the model was made using TensorFlow, evaluating the classification of grooming for a text (dialogue) prepared asynchronously in a controlled environment according to methodologies, techniques, frameworks and multiple proposed techniques with his development described. The model predicts a high number of false positives, therefore low precision and F-score, but a high 88.4% accuracy and 0.81 AUROC (Area under the Receiver Operating Characteristic).
CITATION STYLE
Muñoz, F., Isaza, G., & Castillo, L. (2021). Smartsec4cop: Smart cyber-grooming detection using natural language processing and convolutional neural networks. In Advances in Intelligent Systems and Computing (Vol. 1237 AISC, pp. 11–20). Springer. https://doi.org/10.1007/978-3-030-53036-5_2
Mendeley helps you to discover research relevant for your work.