Stock market predictor

In the past, I researched and implemented Artificial Intelligence models destined to predict the trajectory of pedestrians based on their own past trajectory and that of the people surrounding them. Considering the similarity between the data type and the behavioural rationale driving the change in the series, I consider that implementing a LSTM-based solution that incorporates knowledge about collective interactions may bring a new perspective on market models.

Additionally, I intend to enrich the design with a parallel system that will be in charge of collecting users' activity in Twitter in real time. Eventhough at the beginning this part will only monitor the raw activity (e.g., total number of tweets containing a certain word), I intend to incorporate a basic NLP system.