TransactionBuilder: A function to create sequences ID according to the delay...

In MGousseff/r2spmf: Links R 2 the SPMF Library for frequent itemsets and sequences mining

Description

Input is a data frame and affects a sequence ID to the events related to an ID, built on the time elapsed between an event and thi previous event.

Usage

TransactionBuilder(df, ID, event, time, timeFormat, timeMax = "",
  timeMin = 0, timeUnit = "secs", parallel = F)

Arguments

df	is a dataframe
ID	is the ID, for instance of a customer. Within this ID the function will create one or several sequences, according to time between events.
time	is the time of the event
timeFormat	is the format in which the time is encoded
timeMax	if the time elapsed between two events is more than timeMax, a new sequence is started
timeMin	if the time elapsed between two events is less than timeMin, these events are considered to happen together : they belong to the same itemset
parallel	if True the function will use the several cores available to split the building of the sequence. It appears to be useful above 1 million of lines, depending, of course, of the specs of the computer.

Value

a dataframe with three columns. sequence contains all the frequent sequences. support is the number of times this sequence occurs, and frequence is support divided by the total number of sequences

MGousseff/r2spmf documentation built on May 26, 2019, 11:58 p.m.