REx
Rule Extractor
A Logico-Visual Data Analytics Tool

processing selection → pre- and main processing

# PRE- AND MAIN PROCESSING

## Extraction and Visualisation of a Sound (thus Consistent) and Complete Rulebase

1. INPUT DATA (PREGNANT WITH IMPLICATIONS ;-)
 alternative column specifications: separator = (1 character) width = character(s)
2. CELL VALUE FILTERS (ANONYMISING NORMALISATION)
3. PREPROCESSED INPUT DATA (ANONYMISED & NORMALISED)

$$\mathsf{extraction~cost}\ \ =\ \ \max\left\{\ 50,\ \frac{\mathsf{length}}{10^{5}} \cdot \underbrace{\negthickspace \sum_{r=1}^{r=\mathsf{rank}} (\mathsf{width} \cdot \mathsf{cardinality})^{r+1}}_{\sf\text{rule-diagram(s) size}}\ \right\}\ \sf\text{cents (CHF)}$$ $$\begin{array}{r|l} \mathsf{length} & \sf\text{the number of table rows}\\ \mathsf{width} & \sf\text{the number of table columns}\\ \mathsf{cardinality} & \sf\text{the number of filter codes}\\ \mathsf{rank} & \sf\text{the number of antecedent conjuncts in a rule} \end{array}$$ (In the rule $(A_0\land A_1)\rightarrow C,$ the formula $(A_0\land A_1)$ is the antecedent, $A_0$ and $A_1$ are the antecedent conjuncts, and $C$ is the consequent.)