TABLE 1.

Tweet classification criteria for sentiment analysis.

Sentiment	Criterion
Positive	This category includes tweets that accept genome-edited food or government policies, explain the technology scientifically, or elucidate the reason for non-mandatory labeling.
Negative	This category consists of tweets against genome-edited foods and MHLW/CAA policies or those that call for a signature petition of severe regulation. Many of them include the following terms: “scary,” “don’t want to be distributed,” “right not to eat,” “don’t want to buy,” “don’t want to eat,” “dangerous,” etc.
Neutral	This category includes tweets that are neutral on genome-edited food/labeling or are just publicizing government policies, etc.