根据英语、荷兰语和德语句子对未来的意义进行分类。
ftr-classifier的Python项目详细描述
这是robertson等人描述的关键词分析技术的实现。(TKTK)。
It is designed to classify future-time-referring sentences in English, Dutch, and German in terms of whether they use the accepted “future” tense morphemes (i.e. will/shall/be going to (English); zullen/gaan (Dutch); or werden (German)), use the present tense to refer to future time, or use an expression which specifically marks the probability of the utterance, i.e. “It could rain tomorrow”, or “It will probably rain tomorrow.” This classifier is without doubt overfit to the data it is designed to be used on, and will not perform well out-of-sample. For instance, it assumes all sentences it encounters refer to future time, and that most make predictions, (as ooposed to talk about schedules, or intentions). It is designed to be used exclusively with data generated using the Future Time Reference questionnaire described in Robertson et al (TKTK), and available for free to academic researchers, downloadable link to come.
Any interested researchers should feel free to contact myself (Cole Robertson) at cbjrobertson@gmail.com, or open an issue for discussion on the project git repository.