Social media platforms play an important role in facilitating the spread of extremism by influencing people's views, opinions and perceptions. These platforms are increasingly being used by extremist elements to spread propaganda, radicalize and recruit youth. Therefore, research on identifying extremism on social media platforms is necessary to prevent its impact and negative consequences. A comprehensive and comparative study of datasets, classification methods, and screening methods using an online extremism detection tool is essential.

The purpose of this research is to create a system for identifying ISIS and Al-Qaeda flags from images. Dataset augmentation and training were performed using image augmentation using CNN deep learning networks.  A solution has been found to increase the size of the dataset. The complete dataset contains 1,400 images, half of which are the Al-Qaeda flag and half are the ISIS flag. Additionally, 2 convolutional and one fully connected layer were used to recognize ISIS and Al-Qaeda flags. A method for recognizing ISIS and Al-Qaeda flags was proposed.

As the relevance of the work, it should be noted that the CNN convolutional network model was trained using deep learning, that is, based on neural networks. The novelty of the work is that the model obtains the highest value of the accuracy indicator using the new data set using the classification method.

The study is devoted to the study and application of machine learning methods aimed at solving the problems of identifying potentially dangerous information on the Internet.