An R package to perform outlier detection in contingency tables (i.e. categorical data) using decomposable graphical models (DGMs); models for which the underlying association between all variables can be depicted by an undirected graph. molic are designed to work with undirected decomposable graphs returned from fit_graph
in the ess package. Compute-intensive procedures are implemented using Rcpp/C++ for better run-time performance.
You can install the current stable release of the package by using the devtools
package:
devtools::install_github("mlindsk/molic", build_vignettes = FALSE)
- The Outlier Model: The "behind the scenes" of the outlier model.
- Detecting Skin Diseases: An example of using the outlier model to detect skin diseases.
- Outlier Detection in Genetic Data: An example of how to conduct an outlier analysis in genetic data.
library(dplyr)
library(molic)
library(ess) # For the fit_graph function
set.seed(7) # For reproducibility
Psoriasis patients
d <- derma %>%
filter(ES == "psoriasis") %>%
select(-ES) %>%
as_tibble()
Fitting the interaction graph
g <- fit_graph(d, trace = FALSE) # see package ess for details
plot(g, vertex.size = 15)
This plot shows how the variables are 'associated' in the psoriasis class; see ess for more information about fit_graph
. The outlier model exploits this knowledge instead of assuming independence between all variables (which would clearly be a wrong assumption looking at the graph). The graph may look very different for other classes than psoriasis.
We start by fitting an outlier model taking advantage of the fittet graph g
which holds information about the psoriasis patients. The print method prints information about the distribution of the (deviance) test statistic.
m1 <- fit_outlier(d, g)
print(m1)
#>
#> --------------------------------
#> Simulations: 10000
#> Variables: 34
#> Observations: 111
#> Estimated mean: 42.49
#> Estimated variance: 33.29
#> --------------------------------
#> Critical value: 52.80927
#> Alpha: 0.05
#> <outlier, outlier_model, list>
#> --------------------------------
Notice that m1
is of class 'outlier'. This means, that the procedure has tested which observations within the data are outliers. This method is most often just referred to as outlier detection. The outliers, on a 5% significance level, can now be extracted as follows:
outs <- outliers(m1)
douts <- d[which(outs), ]
douts
#> # A tibble: 8 x 34
#> c1 c2 c3 c4 c5 c6 c7 c8 c9 c10 c11 h12 h13
#> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr>
#> 1 2 1 2 2 2 0 0 0 0 0 0 0 1
#> 2 2 2 2 3 3 0 0 0 0 2 0 0 1
#> 3 3 3 2 2 1 0 0 0 0 1 0 0 2
#> 4 1 1 1 1 1 0 0 0 1 1 0 0 0
#> 5 3 2 1 2 2 0 0 0 2 2 0 0 0
#> 6 1 1 1 1 1 0 0 0 2 2 0 0 0
#> 7 2 3 1 2 1 0 0 0 0 0 0 0 0
#> 8 3 2 3 0 0 0 0 0 3 0 0 0 0
#> # … with 21 more variables: h14 <chr>, h15 <chr>, h16 <chr>, h17 <chr>,
#> # h18 <chr>, h19 <chr>, h20 <chr>, h21 <chr>, h22 <chr>, h23 <chr>,
#> # h24 <chr>, h25 <chr>, h26 <chr>, h27 <chr>, h28 <chr>, h29 <chr>,
#> # h30 <chr>, h31 <chr>, h32 <chr>, h33 <chr>, age <chr>
The following plot is the distribution of the test statistic corresponding to the information retrieved using the print method. One can think of a simple t-test, where the distribution of the test statistic is a t-distribution. In order to conclude on the hypothesis, one finds the critical value and verify if the test statistic is greater or less than this.
plot(m1)
Retrieving the observed test statistics for the individual observations:
x1 <- douts[1, ] %>% unlist() # an outlier
x2 <- d[1, ] %>% unlist() # an inliner
dev1 <- deviance(m1, x1) # falls within the critical region in the plot (the red area)
dev2 <- deviance(m1, x2) # falls within the acceptable region in the plot
dev1
#> [1] 58.97452
dev2
#> [1] 51.05233
Retrieving the p-values:
pval(m1, dev1)
#> [1] 0.0091
pval(m1, dev2)
#> [1] 0.0781
An observation from class chronic dermatitis:
z <- derma %>%
filter(ES == "chronic dermatitis") %>%
select(-ES) %>%
slice(1) %>%
unlist()
Test if z is an outlier in class psoriasis:
m2 <- fit_outlier(d, g, z)
print(m2)
#>
#> --------------------------------
#> Simulations: 10000
#> Variables: 34
#> Observations: 112
#> Estimated mean: 43.73
#> Estimated variance: 36.53
#> --------------------------------
#> Critical value: 54.61947
#> Deviance: 77.92978
#> P-value: 0
#> Alpha: 0.05
#> <novelty, outlier_model, list>
#> --------------------------------
plot(m2)
Notice that m2
is of class 'novelty'. The term novelty detection is sometimes used in the litterature when the goal is to verify if a new unseen observation is an outlier in a homogeneous dataset. Retrieving the test statistic and p-value for z
dz <- deviance(m2, z)
pval(m2, dz)
#> [1] 0
If you want to cite the outlier method please use
@article{lindskououtlier,
title={Outlier Detection in Contingency Tables Using Decomposable Graphical Models},
author={Lindskou, Mads and Svante Eriksen, Poul and Tvedebrink, Torben},
journal={Scandinavian Journal of Statistics},
publisher={Wiley Online Library},
doi={10.1111/sjos.12407},
year={2019}
}
If you want to cite the molic package please use
@software{lindskoumolic,
author = {Mads Lindskou},
title = {{molic: An R package for multivariate outlier
detection in contingency tables}},
month = oct,
year = 2019,
publisher = {Journal of Open Source Software},
doi = {10.21105/joss.01665},
url = {https://doi.org/10.21105/joss.01665}
}