Uncertainty Cue Words

Fig. 3. Distributions of hedge words and uncertainty cue words with at least 10 occurrences in S+ or S-.<br>Fig. 5. A network of 435 words and 2119 similarity links, including 61 seed words (* − 1 − *) and 151 confirmed expanded words (* − * − 1) based on the Google News Word2Vec model. Words are colored by 12 modularity groups. The label size of a word is proportional to its PageRank score.<br>Fig. 7. A network of 435 words based on the PubMed word2vec model in 16 similarity clusters, including 61 seed words, 151 accepted candidates, and the remaining words are rejected candidates.<br>图3. S+或S-中出现频次不少于10次的模糊限制语（hedge words）与不确定性提示词（uncertainty cue words）分布<br>图5. 基于谷歌新闻词向量（Google News Word2Vec）模型构建的含435个词汇与2119条相似性关联的网络，其中包含61个种子词（seed words）（*−1−*）与151个经确认的扩展词（*−*−1）。词汇按12个模块化分组着色，词汇标签的字号大小与其PageRank分值呈正比<br>图7. 基于PubMed词向量（PubMed Word2Vec）模型构建的435个词汇的网络，该网络划分为16个相似性簇，包含61个种子词（seed words）、151个已接受候选词，剩余词汇均为被拒候选词。