Research methods of factoring in the popularity / release date of the recommendation (e.g., if the release date was after the comment or the recommendation was incredibly unpopular or old, then it's probably not a legitimate recommendation) to give the algorithm more insight and have it spit out more accurate results. Then, discuss with team which ones might be best to use (e.g., pros and cons of each one)
Research methods of factoring in the popularity / release date of the recommendation (e.g., if the release date was after the comment or the recommendation was incredibly unpopular or old, then it's probably not a legitimate recommendation) to give the algorithm more insight and have it spit out more accurate results. Then, discuss with team which ones might be best to use (e.g., pros and cons of each one)