Example_on_simulated_data"
In ClustImpute: K-Means Clustering with Build-in Missing Data Imputation

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.width = 7,
  fig.height = 5
)

Simulated data with missings

We start by loading R packages used in this example

required_packages <- Hmisc::Cs(psych,ggplot2,ggExtra,tidyr,Hmisc,tictoc,ClusterR,copula,dplyr,corrplot,ClustImpute)
lapply(required_packages, require, character.only = TRUE)

First we create a random dataset with some structure and a few uncorrelated variables

### Random Dataset
set.seed(739)
n <- 7500 # numer of points
nr_other_vars <- 4
mat <- matrix(rnorm(nr_other_vars*n),n,nr_other_vars)
me<-4 # mean
x <- c(rnorm(n/3,me/2,1),rnorm(2*n/3,-me/2,1)) 
y <- c(rnorm(n/3,0,1),rnorm(n/3,me,1),rnorm(n/3,-me,1))
true_clust <- c(rep(1,n/3),rep(2,n/3),rep(3,n/3)) # true clusters
dat <- cbind(mat,x,y)
dat<- as.data.frame(scale(dat)) # scaling
summary(dat)

One can clearly see the three clusters of the randomly generated data:

dat4plot <- dat
dat4plot$true_clust_fct <- factor(true_clust)
p_base <- ggplot(dat4plot,aes(x=x,y=y,color=true_clust_fct)) + geom_point()
ggExtra::ggMarginal(p_base, groupColour = TRUE, groupFill = TRUE)

We create a 20% missings using a custom function

dat_with_miss <- miss_sim(dat,p=.2,seed_nr=120)
summary(dat_with_miss)
mis_ind <- is.na(dat_with_miss) # missing indicator

The correlation matrix of the missing indicator shows that the missings are correlated - thus we are not in a missing completely at random stetting:

corrplot::corrplot(cor(mis_ind),method="number")

Median or random imputation

Clearly, an imputation with the median value does a pretty bad job here. All imputed values lie on either of the two axes thereby completely distorting the marginal distributions:

dat_median_imp <- dat_with_miss
for (j in 1:dim(dat)[2]) {
  dat_median_imp[,j] <- Hmisc::impute(dat_median_imp[,j],fun=median)
}
imp <- factor(pmax(mis_ind[,5],mis_ind[,6]),labels=c("Original","Imputed")) # point is imputed if x or y is imputed
p_median_imp <- ggplot(dat_median_imp) + geom_point(aes(x=x,y=y,color=imp))
ggExtra::ggMarginal(p_median_imp,groupColour = TRUE, groupFill = TRUE)

But also a random imputation is not much better: it creates plenty of points in areas with no data. Note how a simple check of the marginal distributions would not reveal this issue!

dat_random_imp <- dat_with_miss
for (j in 1:dim(dat)[2]) {
  dat_random_imp[,j] <- impute(dat_random_imp[,j],fun="random")
}
imp <- factor(pmax(mis_ind[,5],mis_ind[,6]),labels=c("Original","Imputed")) # point is imputed if x or y is imputed
p_random_imp <- ggplot(dat_random_imp) + geom_point(aes(x=x,y=y,color=imp))
ggExtra::ggMarginal(p_random_imp,groupColour = TRUE, groupFill = TRUE)

A cluster base on random imputation will thus not provide good results (even if we "know" the number of clusters, which is 3 in this case). Note how the marginal distribution for y differs from the first chart of this vignette where we show the true clusters instead of the predicted clusters from a random imputation.

tictoc::tic("Clustering based on random imputation")
cl_compare <- KMeans_arma(data=dat_random_imp,clusters=3,n_iter=100,seed=751)
tictoc::toc()
dat_random_imp$pred <- predict_KMeans(dat_random_imp,cl_compare)
p_random_imp <- ggplot(dat_random_imp) + geom_point(aes(x=x,y=y,color=factor(pred)))
ggExtra::ggMarginal(p_random_imp,groupColour = TRUE, groupFill = TRUE)

Solution: ClustImpute

We'll now use ClustImpute and also measure the run-time. In short, the algorithm follows these steps

It replaces all NAs by random imputation, i.e., for each variable with missings, it draws from the marginal distribution of this variable not taking into account any correlations with other variables
Weights <1 are used to adjust the scale of an observation that was generated in step 1. The weights are calculated by a (linear) weight function that starts near zero and converges to 1 at n_end.
A k-means clustering is performed with a number of c_steps steps starting with a random initialization.
The values from step 2 are replaced by new draws conditionally on the assign cluster from step 3.
Steps 2-4 are repeated nr_iter times in total. The k-means clustering in step 3 uses the previous cluster centroids for initialization.
After the last draws a final k-means clustering is performed.

The intuition is that observation should be clustered with other observations mainly based on their observed values, while the resulting clusters provide donors for the missing value imputation, so that subsequently all variables can be used for the clustering.

nr_iter <- 10 # iterations of procedure
n_end <- 10 # step until convergence of weight function to 1
nr_cluster <- 3 # number of clusters
c_steps <- 50 # numer of cluster steps per iteration
tictoc::tic("Run ClustImpute")
res <- ClustImpute(dat_with_miss,nr_cluster=nr_cluster, nr_iter=nr_iter, c_steps=c_steps, n_end=n_end) 
tictoc::toc()

To get an overview of what ClustImpute you may run the following commands:

res
summary(res)
attributes(res)

We'll first look at the complete data and clustering results. Quite obviously, it gives better results than median / random imputation.

p_clustimpute <- ggplot(res$complete_data,aes(x,y,color=factor(res$clusters))) + geom_point()
ggExtra::ggMarginal(p_clustimpute,groupColour = TRUE, groupFill = TRUE)

Marginal distributions by cluster and feature

In this toy example we already know that we only have to focus on the "real" features x and y, but this will not be the case in practice. Therefore one can also plot the distributions by cluster and feature.The orange bars are showing the cluster centroids. The plots for x and y are in line what we have seen above, and there is not much difference in the noise features V1 to V4 as one would expect.

plot(res)+xlim(-2.5,2.5)

Alternatively once can also visualize the marginal distributions with a box-plot.

plot(res, type="box")

Traceplot of mean and variance

Packages like MICE compute a traceplot of mean and variance for various chain. Here we only have a single realization and thus re-run ClustImpute with various seeds to obtain different realizations.

res2 <- ClustImpute(dat_with_miss,nr_cluster=nr_cluster, nr_iter=nr_iter, c_steps=c_steps, n_end=n_end,seed_nr = 2)
res3 <- ClustImpute(dat_with_miss,nr_cluster=nr_cluster, nr_iter=nr_iter, c_steps=c_steps, n_end=n_end,seed_nr = 3)
mean_all <- rbind(res$imp_values_mean,res2$imp_values_mean,res3$imp_values_mean)
sd_all <- rbind(res$imp_values_sd,res2$imp_values_sd,res3$imp_values_sd)

mean_all <- cbind(mean_all,seed=rep(c(150519,2,3),each=11))
sd_all <- cbind(sd_all,seed=rep(c(150519,2,3),each=11))

The realizations mix nicely with each other.

ggplot(as.data.frame(mean_all)) + geom_line(aes(x=iter,y=V1,color=factor(seed))) + ggtitle("Mean")
ggplot(as.data.frame(sd_all)) + geom_line(aes(x=iter,y=V1,color=factor(seed))) + ggtitle("Std. dev.")

Quality of imputation and cluster results

External validation: rand index

Below we compare the rand index between true and fitted cluster assignment. For all cases we obtain

external_validation(true_clust, res$clusters)

It is considerably lower for random imputation:

class(dat_random_imp$pred) <- "numeric"
external_validation(true_clust, dat_random_imp$pred)

Of course, it is much higher on the (small number of) complete cases.

## complete cases
idx <- which(complete.cases(dat_with_miss)==TRUE)
sprintf("Number of complete cases is %s",length(idx))
sprintf("Rand index for this case %s", external_validation(true_clust[idx], res$clusters[idx]))

The function also computes a variety of other stats

external_validation(true_clust, res$clusters,summary_stats = TRUE)

Variance reduction

We assess quality of clusters, we compute the sum of squares within each cluster, sum up this value and compare it by the total sum of squares.

res_var <- var_reduction(res)
res_var$Variance_reduction
res_var$Variance_by_cluster

We se a reduction of about 27\% using only 3 clusters, most strikingly for x and y because that these variables define the subspace of the true clusters.

More clusters will capture the random distribution of the other variables

res <- ClustImpute(dat_with_miss,nr_cluster=10, nr_iter=nr_iter, c_steps=c_steps, n_end=n_end)
res_var <- var_reduction(res)
res_var$Variance_reduction
res_var$Variance_by_cluster

We'll do this exercise for a several values of nr_cluster (and need a helper function for that since X is an argument of ClustImpute)

ClustImpute2 <- function(dataFrame,nr_cluster, nr_iter=10, c_steps=1, wf=default_wf, n_end=10, seed_nr=150519) {
  return(ClustImpute(dataFrame,nr_cluster, nr_iter, c_steps, wf, n_end, seed_nr))
}
res_list <- lapply(X=1:10,FUN=ClustImpute2,dataFrame=dat_with_miss, nr_iter=nr_iter, c_steps=c_steps, n_end=n_end)

..and put the variances by cluster in a table

tmp <- var_reduction(res_list[[1]])
var_by_clust <- tmp$Variance_by_cluster
for (k in 2:10) {
  tmp <- var_reduction(res_list[[k]])
  var_by_clust <- rbind(var_by_clust,tmp$Variance_by_cluster)
}
var_by_clust$nr_clusters <- 1:10

While there is a rather gradual improvement for the other variables, x and y have a minimum showing optimality for these variables. Such a plot clearly indicates that 3 clusters are a good choice for this data set.

data2plot <- tidyr::gather(var_by_clust,key = "variable", value = "variance", -dplyr::one_of("nr_clusters"))
ggplot(data2plot,aes(x=nr_clusters,y=variance,color=variable)) + geom_line() + scale_x_continuous(breaks=1:10)