📄️ 4ft-Miner Task
The 4ft-Miner finds association rules of the form If A then S — where A is the antecedent and S is the succedent — that meet a minimum probability threshold and are valid for at least a given number of records. It is the core GUHA procedure for enhanced association rule mining.
📄️ SD4ft-Miner Task
The SD4ft-Miner is the subgroup discovery variant of the 4ft-Miner. Its core idea is: find circumstances under which the probability of a rule differs significantly between two groups. For example, you can check whether the probability of a fatal accident differs between young and older drivers under the same road conditions.
📄️ CF-Miner Task
The CF-Miner (Category Fishery Miner) looks for interesting histograms — it finds conditions under which the distribution of a target categorical attribute shifts in a notable way. For example, you can find circumstances under which accident severity is rising or falling compared to the overall dataset distribution.
📄️ UIC-Miner Task
The UIC-Miner (Uplift in Categories Miner) finds conditions under which rare categories in a target attribute are proportionally boosted. It is designed for situations where a valuable but infrequent outcome — such as a fatal accident, a high-value sale, or a rare disease — needs to be identified and amplified despite its low base occurrence.