An Origin Search System for Trend Expression with Estimating Base Form
There are numerous trend expressions on the Web, which often include additional meaning from their origin. Understanding the expressions in conversation is important in terms of using social networking services. Trend expressions are often used with transforming, therefore it is difficult to identify their base form. In this paper, we implemented an origin search system trend expression with estimating its base form. We use Levenshtein distance, Jaccard distance and Dice distance for estimating the base form. We collected trend expressions and conducted an evaluate experiment the performance of each method.