How to perform a kfino outlier detection
B. Cloez & I. Sanchez
mai 10, 2022
library(kfino)
library(dplyr)
#>
#> Attachement du package : 'dplyr'
#> Les objets suivants sont masqués depuis 'package:stats':
#>
#> filter, lag
#> Les objets suivants sont masqués depuis 'package:base':
#>
#> intersect, setdiff, setequal, union
library(ggplot2)This vignette describes how to use the kfino algorithm on time courses in order to detect impulse noised outliers and predict the parameter of interest.
RAJOUTER DU TEXTE
1 Description of the spring1 dataset
blablabla
data(spring1)
# Dimension of this dataset
dim(spring1)
#> [1] 203 5
head(spring1)
#> # A tibble: 6 x 5
#> # Groups: IDE [1]
#> Poids Date IDE Day dateNum
#> <dbl> <dttm> <chr> <dttm> <dbl>
#> 1 28.6 2017-05-24 00:00:00 250016286863027 2017-05-24 16:34:00 0.469
#> 2 45 2017-05-24 00:00:00 250016286863027 2017-05-24 19:24:00 0.587
#> 3 25 2017-05-25 00:00:00 250016286863027 2017-05-25 05:25:00 1.00
#> 4 43 2017-05-25 00:00:00 250016286863027 2017-05-25 05:45:00 1.02
#> 5 23.4 2017-05-25 00:00:00 250016286863027 2017-05-25 05:58:00 1.03
#> 6 0 2017-05-25 00:00:00 250016286863027 2017-05-25 09:30:00 1.172 Kfino algorithm on the spring1 dataset
2.1 Initial parameters not optimized
# --- Without Optimisation on initial parameters
param2<-list(m0=41,
mm=45,
pp=0.5,
aa=0.001,
expertMin=30,
expertMax=75,
sigma2_m0=1,
sigma2_mm=0.05,
sigma2_pp=5,
K=2,
seqp=seq(0.5,0.7,0.1))
resu2<-kfino_fit(datain=spring1,
Tvar="dateNum",Yvar="Poids",
param=param2,
doOptim=FALSE)
#> [1] "-------:"
#> [1] "No optimisation of initial parameters:"
#> [1] "Used parameters: "
#> [1] 41.0 0.5 45.0
# flags are qualitative
kfino_plot(resuin=resu2,typeG="quali",
Tvar="Day",Yvar="Poids",Ident="IDE")
# flags are quantitative
kfino_plot(resuin=resu2,typeG="quanti",
Tvar="Day",Yvar="Poids",Ident="IDE")
2.2 Initial parameters optimized
# --- With Optimisation on initial parameters
param1<-list(m0=NULL,
mm=NULL,
pp=NULL,
aa=0.001,
expertMin=30,
expertMax=75,
sigma2_m0=1,
sigma2_mm=0.05,
sigma2_pp=5,
K=2,
seqp=seq(0.5,0.7,0.1))
resu1<-kfino_fit(datain=spring1,
Tvar="dateNum",Yvar="Poids",
doOptim=TRUE,param=param1)
#> [1] "-------:"
#> [1] "Optimisation of initial parameters with ML method - result:"
#> [1] "no sub-sampling performed:"
#> range m0: 40.1 43
#> initial m0opt: 41.3
#> initial mmopt: 46.3
#> [1] "Optimised parameters: "
#> Optimized m0: 42.1
#> Optimized mm: 61.1
#> Optimized pp: 0.7
#> [1] "-------:"
# flags are qualitative
kfino_plot(resuin=resu1,typeG="quali",
Tvar="Day",Yvar="Poids",Ident="IDE")
# flags are quantitative
kfino_plot(resuin=resu1,typeG="quanti",
Tvar="Day",Yvar="Poids",Ident="IDE")
2.2.1 Prediction of the weight on the cleaned dataset
kfino_plot(resuin=resu1,typeG="prediction",
Tvar="Day",Yvar="Poids",Ident="IDE")
3 Description of the merinos1 dataset
ewelambs, expected weights between 10 and 45 kg.
data(merinos1)
# Dimension of this dataset
dim(merinos1)
#> [1] 397 5
head(merinos1)
#> # A tibble: 6 x 5
#> # Groups: IDE [1]
#> Poids Date IDE Day dateNum
#> <dbl> <dttm> <chr> <dttm> <dbl>
#> 1 0 2021-01-27 00:00:00 250017033503004 2021-01-27 08:31:00 0.353
#> 2 0 2021-01-27 00:00:00 250017033503004 2021-01-27 08:31:00 0.353
#> 3 0 2021-01-27 00:00:00 250017033503004 2021-01-27 08:31:00 0.353
#> 4 0 2021-01-27 00:00:00 250017033503004 2021-01-27 08:31:00 0.353
#> 5 0 2021-01-27 00:00:00 250017033503004 2021-01-27 08:31:00 0.353
#> 6 40.2 2021-01-27 00:00:00 250017033503004 2021-01-27 12:37:00 0.5244 Kfino algorithm on the merinos1 dataset
4.1 Initial parameters optimized
# --- With Optimisation on initial parameters
param3<-list(m0=NULL,
mm=NULL,
pp=NULL,
aa=0.001,
expertMin=10,
expertMax=45,
sigma2_m0=1,
sigma2_mm=0.05,
sigma2_pp=5,
K=2,
seqp=seq(0.5,0.7,0.1))
resu3<-kfino_fit(datain=merinos2,
Tvar="dateNum",Yvar="Poids",
doOptim=TRUE,param=param3)
#> [1] "-------:"
#> [1] "Optimisation of initial parameters with ML method - result:"
#> [1] "no sub-sampling performed:"
#> range m0: 29 41.8
#> initial m0opt: 36.8
#> initial mmopt: 37.6
#> [1] "Optimised parameters: "
#> Optimized m0: 29
#> Optimized mm: 48
#> Optimized pp: 0.7
#> [1] "-------:"
# flags are qualitative
kfino_plot(resuin=resu3,typeG="quali",
Tvar="Day",Yvar="Poids",Ident="IDE")
# flags are quantitative
kfino_plot(resuin=resu3,typeG="quanti",
Tvar="Day",Yvar="Poids",Ident="IDE")
4.1.1 Prediction of the weight on the cleaned dataset
kfino_plot(resuin=resu3,typeG="prediction",
Tvar="Day",Yvar="Poids",Ident="IDE")
5 Session informations
#> R version 4.1.3 (2022-03-10)
#> Platform: x86_64-w64-mingw32/x64 (64-bit)
#> Running under: Windows 10 x64 (build 19044)
#>
#> Matrix products: default
#>
#> locale:
#> [1] LC_COLLATE=French_France.1252 LC_CTYPE=French_France.1252
#> [3] LC_MONETARY=French_France.1252 LC_NUMERIC=C
#> [5] LC_TIME=French_France.1252
#>
#> attached base packages:
#> [1] stats graphics grDevices utils datasets methods base
#>
#> other attached packages:
#> [1] ggplot2_3.3.6 dplyr_1.0.8 kfino_1.0.0
#>
#> loaded via a namespace (and not attached):
#> [1] tidyselect_1.1.2 xfun_0.30 bslib_0.3.1 remotes_2.4.2
#> [5] purrr_0.3.4 colorspace_2.0-3 vctrs_0.4.0 generics_0.1.2
#> [9] testthat_3.1.3 usethis_2.1.5 htmltools_0.5.2 yaml_2.3.5
#> [13] utf8_1.2.2 rlang_1.0.2 pkgbuild_1.3.1 pillar_1.7.0
#> [17] jquerylib_0.1.4 glue_1.6.2 withr_2.5.0 DBI_1.1.2
#> [21] sessioninfo_1.2.2 lifecycle_1.0.1 stringr_1.4.0 munsell_0.5.0
#> [25] gtable_0.3.0 devtools_2.4.3 evaluate_0.15 memoise_2.0.1
#> [29] labeling_0.4.2 knitr_1.39 callr_3.7.0 fastmap_1.1.0
#> [33] ps_1.6.0 fansi_1.0.3 highr_0.9 scales_1.2.0
#> [37] cachem_1.0.6 desc_1.4.1 pkgload_1.2.4 jsonlite_1.8.0
#> [41] farver_2.1.0 fs_1.5.2 brio_1.1.3 digest_0.6.29
#> [45] stringi_1.7.6 processx_3.5.3 rprojroot_2.0.3 grid_4.1.3
#> [49] cli_3.2.0 tools_4.1.3 magrittr_2.0.3 sass_0.4.1
#> [53] tibble_3.1.6 crayon_1.5.1 pkgconfig_2.0.3 ellipsis_0.3.2
#> [57] prettyunits_1.1.1 assertthat_0.2.1 rmarkdown_2.14 rstudioapi_0.13
#> [61] R6_2.5.1 compiler_4.1.3