简体   繁体   English

IBM Watson Assistant:我应该如何对实体进行分组?

[英]IBM Watson Assistant: How should I group the entities?

I am creating a chatbot to ask some certain questions about the business of a list of stores of many companies in the countries as in the image.我正在创建一个聊天机器人,以询问有关国家/地区许多公司的商店列表业务的某些问题,如图所示。 I am struggling with the fact that how should I group the entities for this case:我正在努力解决我应该如何对这种情况下的实体进行分组的事实:

When the user inputs: Wonder Group, I want to check whether it has any store in California or not.当用户输入:Wonder Group 时,我想检查它在加利福尼亚是否有商店。 -> If Yes, then I will ask for the name of the store -> Check if the store is available in California -> If yes, then answer: OK, Store A, // If no, then say "Sorry, there is no store under this name in California!" -> 如果是,那么我会询问商店的名称 -> 检查该商店是否在加利福尼亚有售 -> 如果是,则回答:OK, Store A, // 如果没有,则说“对不起,有在加利福尼亚没有以这个名字命名的商店!”

-> If No, then answer "Sorry, we do not have information as you require.". -> 如果否,则回答“抱歉,我们没有您需要的信息。”。

Also, should I use the multiple conditioned responses or slot here?另外,我应该在这里使用多个条件响应还是插槽?

I have tried many ways, but it does not work.我尝试了很多方法,但它不起作用。

Entity type实体类型

As mentioned in the comments, more detail will help.如评论中所述,更多细节将有所帮助。 However, just going by your image, it seems you may wish to have a number of entity groups.但是,按照您的形象,您似乎可能希望拥有多个实体组。 Locations (Florida, Texas, etc) and the store groups (Wonder group, snowman etc) would be a good start.地点(佛罗里达州、德克萨斯州等)和商店组(Wonder 组、雪人等)将是一个好的开始。 Then depending on the number of stores in each location, you may wish to hard code these, or use an external database with the stores, which could be accessed via an external API call from within assistant.然后根据每个位置的商店数量,您可能希望对这些进行硬编码,或者使用带有商店的外部数据库,可以通过助手内部的外部 API 调用访问该数据库。
This only defines creating entities via the 'Synonym entity' method, you may wish instead to train your assistant to recognise your entities via the 'Annotation-based method', see the Assistant documentation for more info.这仅定义通过“同义词实体”方法创建实体,您可能希望训练您的助手通过“基于注释的方法”识别您的实体, 有关更多信息,请参阅助手文档。

Using contextual entities is typically best for situations where all entity values are not known to the developer.使用上下文实体通常最适合开发人员不知道所有实体值的情况。 In your case, I'm assuming you will know all possible values.就您而言,我假设您将知道所有可能的值。 As Timd said, the Synonym approach may be best for this use case.正如 Timd 所说,Synonym 方法可能最适合这个用例。

Just FYI, to practice annotating, visit https://www.ibm.com/cloud/architecture/demo/try-watson-assistant-contextual-entities/仅供参考,要练习注释,请访问https://www.ibm.com/cloud/architecture/demo/try-watson-assistant-contextual-entities/

For one of my projects, I have an annotated entity called benefitName.对于我的一个项目,我有一个带注释的实体,名为 benefitName。 benefitName represents all medical procedures such as Physical Therapy, Gallbladder Removal, Heart Bypass Surgery, etc. I determined it would not be possible to find a list of all the procedures my bot may see, and that is why I used an annotated approach. benefitName 代表所有医疗程序,例如物理治疗、胆囊切除、心脏搭桥手术等。我确定无法找到我的机器人可能看到的所有程序的列表,这就是我使用注释方法的原因。

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

相关问题 如何将IBM Watson Assistant与Watson Discovery链接 - How can I link IBM Watson Assistant with Watson Discovery IBM Watson Assistant:如果一个实体值存在于多个实体中应该怎么做 - IBM Watson Assistant: What should be done if one entity value exist in multiple entities IBM Watson Assistant:如何过滤“其他”消息? - IBM Watson Assistant: How can I filter messages for “anything else”? IBM Watson Assistant:如何将图像作为聊天机器人响应发送 - IBM Watson Assistant: How can I send an image as chatbot response 如何在IBM Watson Assistant中插入一个复选框? - How do i insert a checkbox in the IBM Watson Assistant? 如何在 IBM Watson Assistant 的意图中查看用户示例? - How can I see user examples in Intents in IBM Watson Assistant? IBM Watson Assistant:如何使用系统实体sys-number仅获取少量数字 - IBM Watson Assistant: How to get just few numbers with System entities sys-number 如何将IBM Watson Assistant聊天机器人添加到网页上? - How to add a IBM Watson Assistant chatbot on to a webpage? Flutter 中的 IBM Watson Assistant:如何显示选项? - IBM Watson Assistant in Flutter: How to show options? IBM Watson Assistant:用于创建上下文实体缺失的注释选项卡 - IBM Watson Assistant: Annotations tab for creating contextual entities missing
 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM