UIC-Miner Results
The UIC-Miner (Uplift in Categories Miner) is designed to find conditions under which rare categories in a target attribute are proportionally boosted. It is especially valuable for imbalanced datasets where a minority category — such as a rare outcome or a high-value event — is difficult to isolate with standard association rule mining.
Run Summary
At the top of the page, three execution statistics are shown:
| Field | Description |
|---|---|
| Status | Whether the run completed successfully |
| Rules Found | Total number of rules discovered |
| Execution Time | How long the mining process took |
The Run Configuration Details section is collapsible and shows the full task configuration used for this run.
Discovered Rules
The left panel lists all discovered rules in a scrollable table. Each rule is displayed as a colour-coded antecedent expression pointing to the target attribute, with a * symbol indicating the uplift target category. Click any rule to load its detail in the right panel.
Rule Detail
Selecting a rule updates the Rule Detail panel on the right with the following components:
Target Summary
Shows the target attribute and the full list of categories within it (e.g. Fatal, Serious, Slight).
Histogram (Rule)
A histogram showing the distribution of the target attribute's categories among records that match the rule's antecedent. This lets you see how the category distribution shifts within the subgroup defined by the rule.
Histogram (Background)
A histogram showing the distribution of the target attribute across the entire dataset. Comparing this with the Rule histogram makes it immediately visible how much the antecedent boosts the rare category relative to the baseline.
For example, if Fatal accidents make up 1.9% of the full dataset but 9.7% within the matched subgroup, the rule has identified a strong uplift condition for that rare category.
Exporting Results
Use the Export data button in the top right to download the full result set. The Show Raw Output button displays the raw JSON response from the CleverMiner engine.