アクターネットワーク理論から単語埋め込みモデルを用いたTwitter分析を組み立てる

「毒親」データを事例に

抄録

On Twitter, there are posts by a variety of people who do not often appear in the timelines of ordinary users, and there are a certain number of people who continue to post for their own sake, to connect with others who have had similar experiences, and to communicate with society. In this paper, we obtain data of a certain size from these people, create a word embedding model, and perform clustering. By interpreting these posts from the actor-network-theory, we will clarify what kind of validity there is in them.