Monitoring the quality of models in the context of the complexity of banking regulation and changing input data
Relevance of the project
Today, regulation of large industrial companies, banks, ecosystems, and stock exchanges does not take into account feedback effects, including responses to news background, including the possibility of collective collusion of market participants through social media. Oversight and regulatory offences only relate to actually detected events without taking into account the high probability of their occurrence. Only the application of modern AI technologies, including for analysing social media resources, can eliminate these shortcomings.
Project goal
is to create financial regulatory tools with the help of AI technologies, based on scenario-based feedback modelling and taking into account a large number of explicitly fixed and implicitly manifested factors.
Project tasks:
To develop new modelling and forecasting methods
Such methods will have the possibility of subsequent adjustment of supervision and regulation.
Create modern algorithms for processing news feeds, mass media, and social platforms
This is necessary to build separate indicators of news significance, perceptions of both economic events and the tone of the news background and the reaction of other market participants to it (as an example - perceived inflation by a certain group of the population, anxiety about the financial stability of this or that company, willingness to invest in certain assets of the investment market).
To build an interconnected line of sentiment indices
This allows us to take into account in modelling and forecasting not only the direct impact of news and events, but also the reverse, indirect effects.
