Each tweet may be liked/disliked. This should affect the weighting in some way when the recommender system is weighing each individual tweet. So far, it has been decided to simply lower the weight of each term that appears in that tweet. Other possible solutions involve expanding the size of the term frequency document so that it holds more terms that allow for more accurate weighing of tweets, but this could mean including meaningless words in the term frequency document.
Each tweet may be liked/disliked. This should affect the weighting in some way when the recommender system is weighing each individual tweet. So far, it has been decided to simply lower the weight of each term that appears in that tweet. Other possible solutions involve expanding the size of the term frequency document so that it holds more terms that allow for more accurate weighing of tweets, but this could mean including meaningless words in the term frequency document.