Stemming is another practice usually utilized in sentiment analysis. It is the process of reducing words to their root. This lets us handle words that originate from the same root as a single unigram. For example, love, loving, and loved will be all handled as the same unigram, love.