Fascinating World of Neural Nets  
Neural Network Forums
Home PageHome Page : General Neural Networks Topics : Predictive model
  You are currently not logged in. You can view the forums, but cannot post messages. | HOME | Log In | Register | Search | Help
Post a Reply on This Topic Post a Reply on This Topic

Author Topic: Predictive model
cciasyd Posted: 09-Oct-11 18:43
  Edit Edit
Email the Author Mail   View Author's Profile Profile  
We are doing gene expression studies for a few patients. We are also aware of two parameters of each patient.

Although we welcome any other possible ideas, one idea has been to explore the possibility of feeding the samples as input to Feed Forward neural network and train and predict ATLEAST one of the (two) parameters.

The problem is that we are considering 1000 genes for each patient. A partial table indicating a few top genes (out of 1000) and their respective expression values for each patient is given below.

Gene | PatientA PatientAa PatientB PatientC PatientCc PatientE
HS.57x | 6.7742 6.266 6.4349 6.3531 6.502 6.3626
FCGR2B | 7.7461 6.76 6.2499 7.4245 7.1892 8.1365
TRIM44 | 10.4981 10.0555 10.6009 10.5247 10.4809 10.5963
LOC4x | 6.1766 6.4808 6.3642 6.0325 6.2827 6.3011
DGAT2Lx | 5.9597 6.1936 6.0087 5.8458 5.9456 6.0554

Could you advise on whether it is possible, and if so, some way(s) regarding how to make a predictive model (not necessarily using only Feed-forward NN) out of the above information.

Thank you.

Post a Reply on This Topic Post a Reply on This Topic

Copyright © 2001-2003 Pejman Makhfi. All rights Reserved.