Cineplex Brampton Ticket Prices

Association Rules - Transactional Data

The mining association rules is one of the flagship applications of data mining. The idea is update patterns, as co-occurrences in the database. The emblematic example is the analysis of receipts from supermarkets: they want to discover the rules of behavior such as "if the customer bought diapers and wipes, it will buy milk for growth." In which case it may be appropriate to the proper rays in the same area of the store (this is the case with regard to the supermarket I frequent usually). The "if" the rule is called "history", the "so what" is "therefore." It

is possible to find co-occurrences in the individual tables - variables that are manipulated with the usual data mining software. But often, especially through the induction of association rules, data can be in the form of a transactional basis. If we take the example of the analysis of receipts, we have a list of products by cart.

This data representation is quite natural in view of the problem we want to capture. It also has the advantage of being more compact since only the products listed are actually observed in each cart. We need not concern ourselves with products that are not, especially since they can be very numerous if one refers to the number of items that can offer a brand from supermarkets.

As far as this mode of description is natural, it turns out that many programs do not know apprehend directly. We observe curiously a real division between vocational and tools to those from academia. The first most of them can handle this file type. This is the case of software SPAD 7.3 and SAS Enterprise Miner 4.3 we study in this tutorial. The latter, however, require a prior transformation of data to work. We use a VBA macro running in Excel to transform our data base "individuals - Variable Bit suitable for treatment under Tanagra 1.4.37 and 2.2.2 Knime . Attention, we must respect the original specifications, ie focus only on rules indicating the simultaneous presence of products in shopping carts. There is no question, following a coding 'present - absent "poorly controlled, to produce rules highlighting the simultaneous absence of certain products. This may be interesting in some cases, but this is not the purpose of our analysis.

Keywords: association rule, association rules, SPAD 7.3, lock em 4.3, 2.2.2 Knime, filtering rules, lift
Components: A PRIORI
Tutorial: fr_Tanagra_Assoc_Rule_Transactions. pdf
Data : assoc_rule_transactions.zip
References:
Wikipedia, "Association rule learning "