The computation of features and metrics is derived in
- the estimation of Counts, where only the occurrences of single properties focused and
- Metrics, that are partially computed by Counts in relation of document and corpus size. For this step, micro statistics (document level) and macro statistics (corpus level) statistics are supported.
The computation is configurable by a choose-able list for an individual mode or a default-mode that computes all features.
The output of this process is stored under a configured directory featutres and summary.
For details of the metrics, look into the following detailed documentation:
Installation | Input & Data | Functionality & Tasks | Feature Hub | Summarization | Comparison | Aggregation | Config & Run