Search items to browse rules linked to them. Then define your metric thresholds and orders to filter the results.
The definition of an item was set within the Dataiku Application. It could be a product, a product category, a brand, or any characteristic linked to the products purchased in transactions. This application will let you identify relationships found between the items you sell. See the "What is an association rule ?" helper section for more.
An itemset is a set of one or multiple items that were purchased together within your transactions.
An association rule is a relationship between itemsets. The rule {Bacon, Onion} => {Chicken} means "If people buy bacon and an onion, they will also buy chicken". Here we can see that the rule is divided into two parts: a left-hand side and a right-hand side. The left-hand side is composed of the antecedent itemsets: the ones that TRIGGER an association rule. The right-hand side is composed of the consequent itemsets: the ones that are rule OUTCOMES. An association rule can be interpreted with rules metrics. See the "What is a rule metric ?" helper section for more.
The definition of a rule scope was set within the Dataiku Application. Association rules are learned on your transaction history. A rule scope focuses the rules learning process on information related to the transaction context. For example, having transactions involving multiple countries would enable the solution to learn rules dedicated to each of these countries.
The rule {Bacon, Onion} => {Chicken} means "If people buy bacon, they will also buy chicken", but it is nothing more than a raw relationship between an antecedent and a consequent itemset. It doesn't tells us how frequently that rule will occur in transactions (rule Frequency), if we can trust it (rule Confidence), if we would have an advantage to use it (rule Lift) as is, or the strength of the relationship between its itemsets (rule Conviction). Rules metrics bring that information.
Confidence is a rule metric measuring the likelihood to purchase the consequent (or OUTCOME) itemset knowing that the antecedent (or TRIGGER) itemset is also purchased. Confidence is in [0, 1]. The higher it is the more confident you can trust the rule. See also the "What is an association rule ?" and "What is a rule metric ?" helper sections.
Lift is a rule metric measuring how much using a rule is better than relying only on its consequent (or OUTCOME) likelihood taken independently. Lift is in [0, +inf]. A lift below or equal to 1 means that using the rule is not better than using the likelihood to purchase the consequent (or OUTCOME) itemset. The higher the lift, the more powerful the rule. See also the "What is an association rule ?" and "What is a rule metric ?" helper sections.
Conviction is a rule metric measuring the strength of the dependency of the consequent (or OUTCOME) itemset to the antecedent (or TRIGGER) itemset. Conviction is in [0, +inf]. A Conviction below or equal to 1 means that the consequent (or OUTCOME) is independent of the antecedent (or TRIGGER) itemsets. The higher the conviction, the stronger the dependency between the consequent (or OUTCOME) and the antecedent (or TRIGGER) itemsets. See also the "What is an association rule ?" and "What is a rule metric ?" helper sections.