I have a two-dimensional dataset and I calculate a digit for a new datapoint membership degree on that dataset Want to For this purpose, I was thinking about using the subscription function used in Fuzzy C, which means clustering, which is as follows:
I have implemented this function in R and its code is as follows:
subscription function & lt; - Function (Datapoint, Class, Classcenter, Fusifier) {C = NAO (Classcentre) Number = Degree (RBIDID, Classcenter [Class,]) Val = 0 (for 1: C) {Nominator = Dist (RBIDID Data Point), Classics [k] Val = Val + ((number / selector) ^ (2 / (Fusifier-1)) Answer = 1 / Well Return (Answer)}
Now, when I subscribe to it, a datapoint works, fusiform value equals 2 With that, I was hoping to get the membership function value, which is located between 0 and 1. However, my value can go well over 100. I do not understand what's wrong with me here Any clue?
PS: Although it is classified as a separate question, but if you have alternative ideas for any other subscription function that I can use, Comment on posts
Edit: Information on dataset:
I am generating datasets artificially 2000 datapoints of two dimensions that all arise from a random normal distribution Are there. Each of 500 out of the following has four sets: 15,0 0,15 8,35 15,20 then there is a variance for all the datapoints which is generated 1.
Comments
Post a Comment