Converting TSE data to STS with TraMineRextras - type-conversion

Question

I have a data set which is formatted like TSE:
Event is a code of the mode of entry (1 is export, 2 to 5 is partenariat, 6 to 7 is Acq.Mino, 8 is Acq_Majo). Country is not used.
I would like to convert this table to a sequence object, and I understood that the TSE_to_STS function would help me to do so. However it seems that the output is not as I expect.
I ran the code:
seq.sts<-TSE_to_STS(seq.tse.data,id = 1, timestamp = 2, event = 3, stm =NULL, tmin = 1935,tmax = 2018, firstState = "None").
the output is for company 7:
> seq.sts[7,]
My first surprise is that the event happened in year n and are coded in year n+1 (in red in the above output). My second surprise is that in 2013 and in 2015 I lose information about the second entry in Thailand (in yellow in the above output).Third, I don’t understand why the states are added: why do we have in 2012 Acq_Mino.Acq_Majo and not only Acq_Mino?

Answer 1

The TSE_to_STS function is intended to convert time-stamped event sequences into state sequences. This is done by defining the state in which we are after the event occurs. With “stm = NULL”, the state after each event is characterized as the combination of the current state and the event.
Your example is not reproducible because:
We do not know the re-labeling you applied to your events coded from 1 to 8, and
Seems that you mixed the timestamp and event arguments in the TSE_to_STS example command.
I illustrate below with a reproducible example (based on your example except for the country column that we do not use).
library(TraMineRextras)
seq.tse.data <- structure(list(
ID = c(1L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 4L, 4L, 4L, 4L,
4L, 4L, 4L, 5L, 5L, 5L, 6L, 6L, 6L, 6L, 6L, 6L,
6L, 6L, 7L, 7L, 7L, 7L, 7L, 7L, 7L, 7L),
Year = c(2008L, 2010L, 2012L, 2007L, 2009L, 2010L, 2012L,
2013L, 1996L, 1997L, 1999L, 2003L, 2006L, 2008L,
2012L, 2007L, 2007L, 2008L, 2003L, 2007L, 2007L,
2009L, 2009L, 2011L, 2014L, 2016L, 2006L, 2009L,
2011L, 2013L, 2013L, 2015L, 2015L, 2016L),
Event = c(8L, 7L, 8L, 5L, 1L, 8L, 8L, 8L, 3L, 3L, 3L,
3L, 3L, 8L, 1L, 8L, 8L, 8L, 7L, 8L, 8L, 8L,
8L, 8L, 8L, 8L, 8L, 8L, 7L, 7L, 6L, 7L, 6L,
8L)),
class = "data.frame", row.names = c(NA, -34L)
)
seq.tse.data[seq.tse.data$ID==7,]
# ID Year Event
# 27 7 2006 8
# 28 7 2009 8
# 29 7 2011 7
# 30 7 2013 7
# 31 7 2013 6
# 32 7 2015 7
# 33 7 2015 6
# 34 7 2016 8
seq.sts <- TSE_to_STS(seq.tse.data,
id = 1, timestamp = 2, event = 3,
stm =NULL, tmin = 1990, tmax = 2018,
firstState = "None")
For the id 7, we get:
seq.sts[7,which(names(seq.sts)=="a2005"):which(names(seq.sts)=="a2018")]
# a2005 a2006 a2007 a2008 a2009 a2010 a2011 a2012 a2013 a2014 a2015 a2016 a2017 a2018
# 7 None None 8 8 8 8 8 8.7 8.7 8.7.6 8.7.6 8.7.6 8.7.6 8.7.6
In this example, for id=7 the first event (8) occurs in 2006. Therefore the state is None until 2006 and switches to 8 afterwards. In 2011, event 7 occurs. So, the next year (2012) we switch to the state 'events 8 and 7 already occurred', which is denoted as 8.7. When 7 occurs again in 2013, it does not change the state because 7 already occurred while the occurrence of event 6 leads to the state 8.7.6. The latter becomes effective the next year, i.e. in 2014.
Different schemes for determining the state resulting after the occurrence of each event can be defined by providing the stm matrix to the TSE_to_STS function. Use the seqe2stm function for defining the stm matrix. (Look at the help page of this latter function for more details.)