Browsing by Author "Severin Dembele"

Filter results by typing the first few letters
Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    A Machine Learning Based Framework forCollecting and Using Social Media for Real-timeTerrorist Attacks Prediction
    (International Conference on Artificial Intelligence and its Applications, 2023) Lossan Bonde; Severin Dembele
    Terrorism has become a global plague causing insecurity and jeopardizing the development of many countries. In the past few years, terrorism has exploded in Burkina Faso, affecting education, national security, health, and the economy. There is a great need for solutions to detect and stop terrorist attacks before they occur. This research project seeks to use Artificial Intelligence (AI) to mine social media and detect probable future terrorist attacks. This article describes the design of a framework, its partial implementation, and an experiment to validate the technique. The system consists of five steps: taking social media as input, converting it to text, validating it, extracting essential information, predicting its class, and storing it in a dataset. The modest size of the manually produced dataset utilized in the original experiment is a key drawback of the work discussed in this research. The modest size proved inconvenient for Deep Learning algorithms, which operate best with massive datasets. When we complete the entire system, inserting increased data from social media into the dataset will resolve this limitation. The other limitation is the partial implementation of the framework, which does not provide a comprehensive picture of the proposed approach. Our future works will address the remainder parts of the proposed framework.
  • No Thumbnail Available
    Item
    A Unified Generative Artificial Intelligence Approach for Converting Social Media Content
    (IEEE Explore, 2024) Lossan Bonde; Severin Dembele
    Social media content is relevant for many applications, including applications that assist in fighting the plague of terrorism through Artificial Intelligence (AI). However, social media content is diverse in its form - text, image, audio, and video. Depending on the nature of the applications, it may be desirable to convert all these forms into a unique format to ease processing. Once the data is converted into text, it is then possible to organize it into structured tabular data to feed Machine Learning (ML) algorithms for real-time terrorist attack detections. This paper explores using the emerging Generative Artificial Intelligence (Gen AI) tools for converting social media content (text, image, audio or video) into text format suitable for applying machine learning algorithms. The methodology of this research consisted of studying existing Gen AI tools, evaluating and selecting the best among those that offer API or code integration to implement a tool for converting all forms of social media content into text. The main limitation of this work is the small size of the datasets used in the tools’ evaluation. The design and implementation of the proposed solution have been completed, and the tool is ready for use and integration into a framework for collecting and analysing social media content to fight against terrorism.
  • No Thumbnail Available
    Item
    High Accuracy Location Information Extraction From Social Network Texts Using Natural Language Processing.
    (International Journal on Natural Language Computing (IJNLC), 2023-08) Lossan Bonde; Severin Dembele
    Terrorism has become a worldwide plague with severe consequences for the development of nations. Besides killing innocent people daily and preventing educational activities from taking place, terrorism is also hindering economic growth. Machine Learning (ML) and Natural Language Processing (NLP) can contribute to fighting terrorism by predicting in real-time future terrorist attacks if accurate data is available. This paper is part of a research project that uses text from social networks to extract necessary information to build an adequate dataset for terrorist attack prediction. We collected a set of 3000 social network texts about terrorism in Burkina Faso and used a subset to experiment with existing NLP solutions. The experiment reveals that existing solutions have poor accuracy for location recognition, which our solution resolves. We will extend the solution to extract dates and action information to achieve the project's goal.