I am currently performing sentiment analysis classification using a dataset that has one dependent variable, which is "polarity," i.e. "negative" or "positive." I have another dataset that has three total columns. One is "positive," the second is "negative," and the third is the "text" column. XGBoost is my classifier. The negative and positive columns have different numerical values for each observation. A tweet, for example, that is being scored for sentiment analysis can have both positive and negative scores associated with it. This is why I wanted to keep both columns.
Does anyone have an example of how I can use two labels with an XGBoost classifier?
Thanks for any assistance in advance.