cdnowSummary: CDNOW repeat transaction data summary

In BTYD: Implementing BTYD Models with the Log Sum Exp Patch

Description

Data representing the purchasing behavior of 2,357 CDNOW customers between January 1997 and June 1998, summarized as a customer-by-time matrix and a vector of cumulative weekly transactions.

Usage

data(cdnowSummary)

Format

A named list of four elements:

• cbs A customer-by-time matrix with four columns: frequency ("x"), recency ("t.x"), length of observation in the calibration period ("T.cal"), and number of transactions in the holdout period ("x.star"). Each row represents a customer.

• cu.tracking A vector containing cumulative transactions for every week in both the calibration and estimating periods (78 weeks total). This vector contains the sum of transactions across all customers.

• est.params A vector containing estimated values for the four Pareto/NBD parameters: r, alpha, s, and beta, in that order. This estimation was made using pnbd.EstimateParameters, and is included here to avoid having to run the relatively time-consuming parameter estimation function in examples.

• m.x A vector containing the average value of each customer's repeat transactions. Used in examples for spend functions.

Details

The customers in this data represent 1/10th of the cohort of customers who made their first transactions with CDNOW in the first quarter of 1997. CDNOW was an online retailer, selling music and related products on the web since 1994.

Source

The data was put together using data conversion functions included in this package. The original event log is included (see cdnowElog).

BTYD documentation built on Nov. 18, 2021, 1:10 a.m.