原书出处
Natural Language Processing in Action, Second Edition
All the previous approaches to semantic analysis failed to take into account information about the similarity between documents. We created topics that were optimal for a generic set of rules. Our unsupervised learning of these models for feature (topic) extraction didn’t have any data about how close the topic vectors should be to each other. We didn’t allow any “feedback” about where the topic vectors
原书定位详情
来源:a19c7933-4eda-4176-aebd-30fe7e1ad292
正文位置:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-4.html
EPUB SHA-256:0e9a35e49ca446e322e5524e53f39fb6bee5b2b23d9cda10ef052aca770b0b36