mhahsler / streammoa Goto Github PK

First of all I would like to thank you for the great StreamMOA R package.
There is something I did not understand from the documentation of the package and would love to sharpen the knowledge.
get_assignment function of any stream algorithm return assignments to each given point or NA if any cluster does not fit unless model and type parameters is 'auto' and than if the model does not fit to any cluster "nn" method will assign its decision.

Below is a quote from the documentation:
״method = "auto" selects model-based assignment if available and otherwise defaults to nearest neighbor assignment. Note that model-based assignment might result in some points not being assigned to any cluster (i.e., an assignment value of NA) which indicates a noise data point.״

If I understand correctly, when I enter model = auto I should not get NA in the results but in fact I do get NA.
assignments <- get_assignment(dStream,points[,0:5],type = "auto",method="auto")
Could you please refine my reason for receiving NAs in this case?

Hi! I want to get the value of the distance between the points to the closest center(before create a new center).

this is the part of my code:

stream <- DSD_ReadCSV(file = "mds40_ay4_ba1.csv", header = TRUE)
data <- read.csv("mds40_ay4_ba1.csv", header = TRUE)
data <- data.matrix(data)
reset_stream(stream)
ct <- DSC_ClusTree(horizon = 180, maxHeight = 8, lambda = NULL, k = NULL)
maxDimension <- 1
eps <- 0.029

layout(matrix(c(1,2,2,3,4,4), nrow = 2, ncol = 3, byrow = TRUE))
Diag1<-0
wasserp2d01 <- numeric(9)
streamProgress <- integer(9)
newvalue<-0
for (i in 1:16) {
update(ct, stream, 400)
centers <- data.matrix(na.omit(get_centers(ct, type = "micro")))

And it only gives me the value of all the centers. Could you give me some advice to change it? Thanks so much!