diff --git a/course_files/clust-intro.Rmd b/course_files/clust-intro.Rmd index eda4a81b22a28481655ee03d40fdf1564c62255a..c2061aac62f08040c66d76d70769a874f3dc0401 100644 --- a/course_files/clust-intro.Rmd +++ b/course_files/clust-intro.Rmd @@ -209,7 +209,7 @@ that are very fast, although not the most accurate approaches. -#### Concensus clustering (more robustness, less computational speed) +#### Consensus clustering (more robustness, less computational speed) ##### __Motivation (Two problems of $K$-means)__: \ - __Problem1:__ sensitive to initial partitions \ @@ -219,7 +219,7 @@ that are very fast, although not the most accurate approaches. __Solution:__ Run $K$-means with a range of $K$'s. -##### __Algorithm of concensus clustering (simpliest version)__: +##### __Algorithm of consensus clustering (simpliest version)__: ```{r, eval = F, highlight = F} for(k in the range of K){ for(each subsample of the data){ @@ -252,7 +252,7 @@ Say we partitioned four data points into 2 clusters. <center>{width=60%}</center> -- __Step2:__ Concensus matrix: \ +- __Step2:__ Consensus matrix: \ Average of all the partitions <center>{width=30%}</center>