Predictive modeling of dynamic online social behavior using dynamic graph neural networks

Purpose: The purpose of the work is to train a Dynamic Graph Neural Network machine learning model that, utilizing sequential representations of the social network in the form of graphs from information extracted from social media (Social Media: i.e. X), in order to be able to distinguish the model when a Social Media trend leads to a social event, such as a march, a protest, a mass aggressive behavior. Obviously, for this to be possible, the model will be fed with the information about which successive social graph representations actually led to some event and which ones did not. So the result will be that by giving the model time-sequenced social graph representations concerning a specific trend in social media, it will be able to recognize the possibility of leading to a social event such as a march in the streets. Finally, from one successive representation to the next, in addition to the links between the nodes that will change, the information for each node can be updated in the form of node embeddings and information such as the location of the user-node and possibly also the possible movement (traffic) around its location. Note 1: I’ve personally progressed quite a bit with the work in question, I’ve formed the workflow, found some involved datasets, and written the body of both the code and the model. I have formed the procedure for the preprocessing of the data, however the main data, for example from X, I have not yet collected and I have not tested the model. Still, the main trunk is well defined and it is understood that I listen to suggestions that improve the result without, however, being distracted from the work by the dementia of the scriptures, the social network and possibly most of the general philosophy behind Dynamic Graph Neural Networks. Personally, due to extraordinary circumstances in my life, I do not know if I will be able to devote all the required time needed to finish the subject I started, although my interest in it remains unshaken and the primary foundation of the work in question has already been built. I have more material to make available, and you have been clarified to provide additional information, as soon as the right person is found to take over and move forward with this project.