Probabilistic Graphical Models

Scenario Analysis on Climate Change, Pandemics (e.g. COVID-19), Credit Risk, Decision-Making under Uncertainty and other risk related topics are difficult domains to analyse. The analysis typically involve large collections of random variables with complex conditional dependency structures. Probabilistic Graphical Models (PGM) are a good fit to model such complexity as they bring together probability theory and graph theory through a fully interpretable structure. Our PGM library is a powerful, flexible framework that enables clients to run their analysis, incorporating expert knowledge.

Data Imputation

Missing data is part of almost every dataset in practice. In several cases we cannot just discard the missing data, instead we need to impute it. Backtesting is an obvious example where data imputation has a lot of value. We employ Machine Learning techniques to provide realistic and fast data imputation.

Equity Markets Sentiment Analysis

Social Media is growing in volume and will inevitably be part of trading sources for pricing and risk management. We have worked on using Artificial Intelligence to provide sentiment analysis on major stocks and indices.

Time Series Forecasting

We employ state of the art Deep Learning techniques to provide time series forecasting for several horizons including daily or higher frequencies. Commodities, financials, and other type of time series can be predicted in a framework that is capable of combining text (e.g. news feed) and numbers (e.g. other input time series). 

Drilling Events Detection

Drilling is one of the most expensive and risky upstream activities. It is important to detect in advance events (e.g. stuck rig in mud, broken rig) in order to act fast and efficiently. We use Natural Language Processing combined with advanced Machine Learning techniques to detect potential downtime in advance. This technique also diagnoses non-productive wells in real time. 

Please reload

© 2020 by Tenokonda Ltd. All rights reserved.