we're using human data from different sources/labs to try and predict the protocol used. We used the already trained model on human data. the predictions were erroneous: e.g. for ATAC dataset of 48, only 4 were predicted correctly. for WES data of 6, 2 were predicted as ChIPseq
We don't know if it would be useful to retrain the model on the data we have to get better results. what do you recommend?
thanks
we're using human data from different sources/labs to try and predict the protocol used. We used the already trained model on human data. the predictions were erroneous: e.g. for ATAC dataset of 48, only 4 were predicted correctly. for WES data of 6, 2 were predicted as ChIPseq
We don't know if it would be useful to retrain the model on the data we have to get better results. what do you recommend?
thanks