Data Mining for Business Analytics in Retail

Data Mining for Business Analytics in Retail

Min Gan (Deakin University, Australia) and Honghua Dai (Deakin University, Australia)
Copyright: © 2014 |Pages: 10
DOI: 10.4018/978-1-4666-5202-6.ch057

Chapter Preview



Large volumes of retail transaction data have been accumulated since the use of point of sale (POS) machines and the application of information technology in retail in the 1970s. The retail transaction data are normally organised in relational databases. As shown in Table 1, each record in the database represents one transaction with four major fields; TID (Transaction ID), transaction time, the details of items contained in the transaction, and Total. Transactions of customers using loyalty cards can be further organised as databases of transaction sequences as shown in Table 2. Each row in Table 2 records transactions of one customer in time order. The first column represents Customer ID. The second column contains the transaction sequence of the customer, where each transaction is distinguished by a transaction ID.

Table 1.
A sample database of retail transactions
TIDTimeItems’ DetailsTotal ($)
ItemUnit PriceQuantitySubtotal ($)
1000000109:15, 12/03/2012Beer$3.582$7.16$31.33
Apple$2.451.8 kg$4.41
1000000209:20, 12/03/2012Sugar$1.261.5kg$1.89$15.73
9999999916:34, 09/11/2012Toothpaste$2.92$4.80$8.05

Key Terms in this Chapter

Market Basket Analysis: A direct application of association rule mining in sales transaction databases, where each record represents a sales transaction and an item is a product. For example, association rule “Coffee?Milk” found in a transaction database indicates an association exists between Coffee and Milk, and reveals that customer who purchase coffee tend to buy milk at the same time.

Classification: Classification is a supervised machine learning technique which predicts the classification label of a given instance.

Association Rule Mining: The process of extracting association rules of the form X?Y from a data set. Association rule X?Y indicates that if X occurs in a record in the data set, then Y is likely to appear in the same record

Maximal-Profit Item Selection: In retail business analysis and optimization, maximal-profit item selection focuses on finding item sets with an aim to obtain maximal profits under constraints in business practices, e.g., the maximal number of items to be selected, the maximal limitation of shelf space and the maximal limitation of budget.

Retail Knowledge Discovery: Discovering regularities by mining data accumulated in the retail industry.

Reliability in Knowledge Discovery: It focuses on three major issues; whether discovered knowledge is reliable and trustable, what factors affect reliability, how reliability can be enhanced or ensured.

Data Mining: Data mining is a technique which aims to discover hidden and significant patterns and regularities from large amounts of data.

Complete Chapter List

Search this Book: