How are the top 5000 features chosen during Data Filtering?

[angavalos]

I have a dataset with more than 5000 features and was wondering: how is the data limited to 5000 features during data filtering?

I’ve seen in the “Can I perform PCA using all features (i.e. not filtered to a max. of 5000)?” post that a chosen ranking method is used. What is this method? If I input my data with over 5000 features and click the “None” option for data filtering, how are those 5000 features chosen?

Thanks

[jeff.xia]

We have provided the details and reasoning in the “Data Filtering” page. Your question is covered in the last sentence (see below)