recursive feature extract in model_summary in random forest model

Hi I tried to see how many feature counts could give me a accuracy prediction above certain threshold in the model_summary.qzv file (qiime2-2019.7), however, the feature count is on the scale of 100, so I am wondering if there is any table data support the pdf file then I could get an accurate number?

Thanks.

Thanks for the suggestion @Meng_Wu! There is not currently a way to download the raw data used to generate that plot, so I have opened an issue to track this — contributions to the source code are always very welcome! :wink: