Thesis: “Reflexions methodologies autour du Napping: vers une integration du comportement du sujet dans l’analyse des donnees de Napping”

Chapter 4 - Illustration of functions cpdm(), plotcpdm(), traject(), and dimtracking()

Author: Tam Minh LE

Supervisor : Francois Husson & Sebastien Le

R version: 3.2.1 - Updated: 17.08.2015

rm(list=ls())
options(max.print=5.5E5)   # max print

Require package ———————————————————

library(FactoMineR)
library(SensoMineR)
library(ggplot2)
library(reshape2)
library(stringr)
library(holos)
library(car)
library(cluster)
library(grDevices)
library(knitr)

Import data ————————————————————-

data(cards.fico)  # final configuration

Functions: cpdm & plotcpdm ———————————————-

cards.cpdm <- cpdm(cards.fico, nb.step=3)
  # names(cards.cpdm)
  # cards.cpdm$Steps.of.all.subjects

plotcpdm(cards.cpdm, subject=1)

  # plotcpdm(cards.cpdm, subject=1, axes=c(1,2))  # s1: 40 subjects
  # plotcpdm(cards.cpdm, subject=41, axes=c(1,2)) # s2: 30 subjects
  # plotcpdm(cards.cpdm, subject=71, axes=c(1,2)) # s3:  5 subjects
  # plotcpdm(cards.cpdm, subject=76, axes=c(1,2)) # s4:  5 subjects
  # plotcpdm(cards.cpdm, subject=81, axes=c(1,2)) # s5: 30 subjects

Temporal data is the cards data

  # #' Require package ---------------------------------------------------------
  # library(FactoMineR)
  # library(SensoMineR)
  # library(ggplot2)
  # library(reshape2)
  # library(stringr)
  # library(car)
  # library(cluster)
  # library(grDevices)

Load data —————————————————————

data(cards.tempo)
cards.tempo[1:32,]   # structure of temporal data
##    subject step stimulus     time coordX coordY
## 1        1    0        M 16:32:28      0      0
## 2        1    0        K 16:32:28      0     80
## 3        1    0        P 16:32:28      0    160
## 4        1    0        D 16:32:28      0    240
## 5        1    0        H 16:32:28      0    320
## 6        1    0        A 16:32:28      0    400
## 7        1    0        G 16:32:28      0    480
## 8        1    0        O 16:32:28      0    560
## 9        1    0        L 16:32:28     80      0
## 10       1    0        N 16:32:28     80     80
## 11       1    0        I 16:32:28     80    160
## 12       1    0        E 16:32:28     80    240
## 13       1    0        J 16:32:28     80    320
## 14       1    0        F 16:32:28     80    400
## 15       1    0        C 16:32:28     80    480
## 16       1    0        B 16:32:28     80    560
## 17       1    1        B 16:32:35    887    326
## 18       1    2        C 16:32:38    847    302
## 19       1    3        J 16:32:42    897    249
## 20       1    4        L 16:32:45    940    317
## 21       1    5        A 16:32:47    818    381
## 22       1    6        D 16:32:50    793    234
## 23       1    7        G 16:32:52    757    329
## 24       1    8        P 16:32:55    865    179
## 25       1    9        O 16:33:00    122    323
## 26       1   10        F 16:33:01    122    424
## 27       1   11        H 16:33:02     55    373
## 28       1   12        E 16:33:03      0    303
## 29       1   13        I 16:33:04     65    268
## 30       1   14        N 16:33:06    124    242
## 31       1   15        K 16:33:09     58    190
## 32       1   16        M 16:33:11    196    301

Function traject() ——————————————————

corr.j1  <- traject(cards.tempo, juge=1, axes=c(1,2))

corr.j1[,1:3]
##         Dim.1  Dim.2  Dim.3
## Step0  0.0029 0.0618 0.0057
## Step1  0.0834 0.0663 0.0449
## Step2  0.1587 0.0891 0.0017
## Step3  0.0533 0.1998 0.0162
## Step4  0.1290 0.2859 0.0696
## Step5  0.2450 0.3446 0.0110
## Step6  0.1506 0.4919 0.0013
## Step7  0.0806 0.6530 0.0380
## Step8  0.0357 0.9509 0.0073
## Step9  0.0428 0.9426 0.0030
## Step10 0.0406 0.9447 0.0026
## Step11 0.0383 0.9489 0.0027
## Step12 0.0442 0.9480 0.0018
## Step13 0.0483 0.9480 0.0009
## Step14 0.0485 0.9436 0.0008
## Step15 0.0515 0.9387 0.0017
## Step16 0.0660 0.9336 0.0012
names(corr.j1)
## NULL

Function dimtracking() ————————————————–

dimtracking(corr.j1, num.dim=1:3)
## Using id as id variables

dimtracking() with learning phase

corr.j1  <- traject(cards.tempo,juge=1, axes=c(1,2), plot.graph=FALSE)
res.sum <- suminfo(cards.tempo)
j1.learning <- res.sum$Step.Time$Step.learning[1] 
p <- dimtracking(corr.j1, num.dim=1:3)
## Using id as id variables
p + geom_vline(xintercept = j1.learning, colour="magenta", linetype="longdash")

******************************** End *************************************