使用StanfordCoreNLP提取两个实体之间的关系 [英] Extract Relationship Between two Entities using StanfordCoreNLP

问题描述

这里也提出了类似的问题，但我找不到相关的答案，所以我再试一次。
我可以使用库获取NER和Dependency树。现在我要找的是我想用实体之间的关系提取实体。例如，flipkart投资了myntra，所以我应该能够将entity1作为flipkart，将entity2作为myntra和投资者作为关系。或类似的结构。我无法获得相同的正确工具。我有些指导家伙，如何实现这个目标？提前致谢

the similar question has been asked here but i cant find any relevant answer of it so i am trying again. I am able to get NER and Dependency tree using the library. Now what i looking for is that i want to extract entities with the relationship between the entities. For example , "flipkart has invested in myntra" so i should be able to get entity1 as "flipkart " and entity2 as "myntra" and "investor" as the relation .or similar kind of structure. I am not able to get correct tool for doing the same. I neee some guidelines guys, how to achieve this?. Thanks in advance

推荐答案

您可以使用斯坦福关系提取工具包。与代码一起发布的模型只有几个关系：Live_In，located_In，OrgBased_In，Work_For和None。要训​​练您自己的模型，请在其网页上查看培训并指定您自己的模型。现有模型是在一个小数据集上训练的，所以如果你添加更多的训练数据，你将会有很长的路要走。

You can train your own model using Stanford relation extraction toolkit. The model released with the code has only a few relations: Live_In, Located_In, OrgBased_In, Work_For, and None. To train your own model, look at "Training and specifying your own model" on its webpage. The existing model is trained on a small dataset so you will go a long way if you add more training data.

如果你想在句子中的两个实体之间提取突出的单词（也就是说，不想将它们分类为一组给定的关系），你也可以查看 ReVerb 。

If you want to extract prominent words between two entities in sentences (that is, do not want to classify them into a given set of relations), you can also look at ReVerb.

这篇关于使用StanfordCoreNLP提取两个实体之间的关系的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！