A big data driven multilevel deep learning framework for predicting terrorist attacks.

Area of Science:

• Computer Science
• Security Studies
• Data Science

Background:

• Terrorism poses a significant threat to human security, causing widespread violence and societal unrest.
• Existing deep machine learning models for predicting terrorist attacks are limited by data handling, accuracy, and adaptability.
• There is a critical need for advanced predictive models capable of processing big data for effective counter-terrorism strategies.

Purpose of the Study:

• To develop an integrated Big Data deep learning-based predictive model for forecasting terrorist attack probabilities.
• To address the limitations of current models in handling large datasets and improving prediction accuracy.
• To provide a tool for law enforcement to anticipate and prevent potential terrorist attacks.

Main Methods:

• A Big Data long short-term memory (LSTM) network was developed, treating terrorist activities as a sequence modeling problem.
• The layered LSTM model is designed for processing large-scale datasets and learning from historical event patterns.
• The model was evaluated using the global terrorism dataset, analyzing performance on key metrics.

Main Results:

• The proposed Big Data LSTM model demonstrated promising accuracy in predicting the probability and location of terrorist attacks.
• The model achieved high performance on standard evaluation metrics, including accuracy, precision, recall, and F1 score.
• Experimental results confirm the model's substantial contribution to predicting attacks at city, country, and regional levels.

Conclusions:

• The developed Big Data deep learning model offers a significant advancement in predicting terrorist attack probabilities and locations.
• Accurate prediction of potential attack sites enables law enforcement to implement effective preventative measures.
• This research provides a valuable tool for enhancing national and global security against terrorism.