ADNI data processing

Hi,

Thanks so much for having your code and data processing code on github!

I am trying to process the ADNI dataset and follow your processing scheme. I noticed that on page 6 of your work here, you mention SUVR being no smaller than 1.08 is used to label a data point as positive (beside it being clinically positive). But below I see you also incorporate SUVR==1.18, and I am not sure where this value comes from, and why T (which I am assuming is the true label of the data) is set to 1 when SUVR>=1.18 but not when SUVR>=1.08?

Self-PU/adni_dataset.py

Lines 189 to 194 in a0e332a

    
           if suvr >= 1.18: 
        
               self.Y[i] = 1 
        
               self.T[i] = 1 
        
           elif suvr >= 1.08: 
        
               self.Y[i] = -1 
        
               self.T[i] = 1

Thank you!

Hello,

Thank you for your interest in our project! I must admit it has been quite a while since we worked on this project, and some details escape me at the moment. I'll try to consult with my collaborators and follow up with you promptly.

Best regards,

Thank you! Looking forward to hearing back from you on this!

	if suvr >= 1.18:
	self.Y[i] = 1
	self.T[i] = 1
	elif suvr >= 1.08:
	self.Y[i] = -1
	self.T[i] = 1