Development of a terrorism threat detection system using natural language processing

Terrorism remains a major global security challenge, and the increasing use of digital communication platforms has made it easier for individuals or groups to spread threatening content and coordinate harmful activities. As a result, there is a growing need for intelligent systems capable of automatically detecting potential threats from large volumes of textual data. This study focuses on the development of a terrorism threat detection system that utilizes Natural Language Processing (NLP) and machine learning techniques to analyze text and identify messages that may contain terrorism-related threats.
The system was developed using a dataset consisting of various text samples categorized as terrorism-related threats, suspicious messages, and normal or non-threatening communication. The dataset underwent several preprocessing steps including text cleaning, tokenization, stop-word removal, and stemming or lemmatization. The processed text was then converted into numerical form using vectorization techniques enabling it to be used by machine learning models. Suitable classification models were trained on a training dataset, while a separate testing dataset was used to evaluate performance.
The evaluation of the system was carried out using performance metrics such as accuracy, precision, recall, F1 score, and Matthews Correlation Coefficient (MCC). The results demonstrated that the model effectively identified patterns associated with terrorism-related threats in textual data, achieving perfect scores across all metrics over three training epochs.
The system showed promising performance in distinguishing between threatening and non-threatening messages, indicating that machine learning and NLP techniques can be effectively applied to support automated threat detection and enhance security monitoring systems.