Unified AI Analytics

Auto-build probabilistic models that enable data scientists to capture complex events. Flek models are semi-supervised models that self-adjust with new data.

AI citizens can query and mine for interesting joint & conditional probability patterns. Flek allows users to tabulate & visualize their search, explain their queries as well as trace their prediction results.

With auto-discovery users can scan their models and extract interesting statistics, associations, rules, bisociations, anomalies, polymalies, influencers or causal relationships.

Prediction is bi-directional with multi-feature targets. Both forward and backward predictions and classification tasks can run without retraining if the input or the target variable(s) change.

Flek provides unique probabilistic algorithms which combine both profile and item based recommendation at the same time. Thus, enabling for more customization and higher conversion rate.

Python Toolkit provides the components needed to develop advanced programs and run complex Algorithmic Probability Programming (APP) tasks.