Corpus-driven macro classification results – the Wilbur corpus.

Classification results – macro-precison/recall/F1– for the Wilbur corpus using top 1% and 5% of the rules and including and excluding the det and prep typed dependencies. Similar to Table 6, the highest F1 scores achieved by a type within a particular category of results are represented by the bolded values.

基于威尔伯语料库（Wilbur corpus）的分类结果——宏精确率（macro-precision）、召回率（recall）与F1值（F1）——，实验采用排名前1%与前5%的规则，并分别包含与排除了限定词（det）与介词（prep）类型的依存关系。与表6类似，在某一类结果范畴内，某一类型所取得的最高F1值以加粗字体标注。