search

justsmart / RecFormer

Code for "Information Recovery-Driven Deep Incomplete Multi-view Clustering Network"
10 stars 1 forks source link

Incomplete data #1

Open lywinaaa opened 1 year ago

lywinaaa commented 1 year ago

How the “percentDel_0.5.mat” is obtained

justsmart commented 1 year ago

All you need to do is generate a matrix of shape n*m with all elements 1 (n for the number of samples, m for the number of views), and follow the rule to set 50% of the 1 to 0. The specific rule is: set the random half of the elements in each column of the matrix to 0, and then check if at least one view is available for each sample, if not, you need to fine-tune the position of the 0 and 1 elements in the matrix until the requirement is met.

justsmart commented 1 year ago

And this generation process requires you to implement using the Matlab or python

lywinaaa commented 1 year ago

而这个生成过程需要你使用Matlab或python来实现

Hello, first of all, thank you very much for your reply, because I am new to this field, so I may not understand what you said, if possible, can you provide your code for constructing incomplete data? Or existing incomplete datasets from your thesis experiments. Finally, I wish you smooth work and good health!

lywinaaa commented 1 year ago

而这个生成过程需要你使用Matlab或python来实现

您好，还有个问题就是您为了减少实验随机性，一般都是执行几次试验

justsmart commented 1 year ago

而这个生成过程需要你使用Matlab或python来实现

您好，还有个问题就是您为了减少实验随机性，一般都是执行几次试验

一般5-10次重复实验