Emphasize certain tokens for classification
I apologies if this has been answered somewhere else. all my search so far has yielded nothing related to my situation.
I would like to be able to emphasize or create more weight to certain keywords/tokens to be applied to my classification.
I currently have a pretty good text mining model that predicts my True Positive classes 86% (Recall) of the time which is acceptable but I would like to improve that by introducing more weight to certain tokens if possible.
The aim of this is to increase the classification of the true positive by making sure the model accounts for those keywords. I am processing the documents using TF-IDF.
Please let me know if you have any questions about my use case and looking forward to hearing from the community.