• United States
IBM?
  • Site map
IBM?
  • Marketplace

  • Close
    Search
  • Sign in
    • Sign in
    • Register
  • IBM Navigation
IBM Developer Answers
  • Spaces
    • Blockchain
    • IBM Cloud platform
    • Internet of Things
    • Predictive Analytics
    • Watson
    • See all spaces
  • Tags
  • Users
  • Badges
  • FAQ
  • Help
Close

Name

Community

  • Learn
  • Develop
  • Connect

Discover IBM

  • ConnectMarketplace
  • Products
  • Services
  • Industries
  • Careers
  • Partners
  • Support
10.190.13.206

Refine your search by using the following advanced search options.

Criteria Usage
Questions with keyword1 or keyword2 keyword1 keyword2
Questions with a mandatory word, e.g. keyword2 keyword1 +keyword2
Questions excluding a word, e.g. keyword2 keyword1 -keyword2
Questions with keyword(s) and a specific tag keyword1 [tag1]
Questions with keyword(s) and either of two or more specific tags keyword1 [tag1] [tag2]
To search for all posts by a user or all posts with a specific tag, start typing and choose from the suggestion list. Do not use a plus or minus sign with a tag, e.g., +[tag1].
  • Ask a question

Watson Natural Language Classifier - need low confidence for questions not in training set

31000134BP gravatar image
Question by kavitha.chittibabu  (48) | Nov 12, 2015 at 01:00 AM watsonnatural-language-classifier

Hi, I am building an appointment booking system in hospital. I need Natural Language Classifier to return a low confidence level for the questions that are not listed in the training set. I had a look at this link where its advised to setup a "notRelevant" classify - with a set of sample questions that do not match any of the normal set.

https://developer.ibm.com/answers/questions/207490/how-to-make-natural-language-classifier-return-low.html

I have tried this and NLC returns "notRelevant" based on the sample sets again (obviously thats the behaviour of NLC - answers based on sample set).

But covering the negative scenario for any domain is not possible. i.e, my domain is appointment booking in hospital. But, if I ask NLC "Tell me world wonders", it returns with a classifier in the training set with higher confidence level.

Is there any solution handling negative / non-domain scenarios?

People who like this

  0
Comment
10 |3000 characters needed characters left characters exceeded
  • Viewable by all users
  • Viewable by moderators
  • Viewable by moderators and the original poster

1 reply

  • Sort: 
110000PNBC gravatar image

Answer by @chughts (11657) | Nov 12, 2015 at 02:44 AM

The hurdle that you are hitting is that if you train the Natural Language Classifier (NLC) to only recognize hospital related statements, then it will see hospital related statements in everything it sees.

If you take a look at the response that you get, you will notice that it gives a confidence level against every classification in the corpus you created, and not only the classification that it ranks with the highest confidence level. You get higher confidence levels the closer NLC can match the statement with a classification statement that it has been trained with.

In your case you might want to nest your classifiers, ie. start off with a simple "relevant", "notrelevant", split then pass it to your hospital classifier, knowing that the statement has a higher confidence level of being relevant.

Comment

People who like this

  0   Share
10 |3000 characters needed characters left characters exceeded
  • Viewable by all users
  • Viewable by moderators
  • Viewable by moderators and the original poster

Follow this question

56 people are following this question.

Answers

Answers & comments

Related questions

Can I use a classifer of ibmwatson-nlc-groundtruth application in my own code? 1 Answer

Natural Language Classifier Conflict with this classifier error 0 Answers

Phrase Length in Watson Natural Language Classifier 4 Answers

How to get questions and classifications from a classifier 0 Answers

NLC and Node-RED, choosing the NLC classifier_id via msg (or other runtime variable) 1 Answer

  • Contact
  • Privacy
  • IBM Developer Terms of use
  • Accessibility
  • Report Abuse
  • Cookie Preferences

Powered by AnswerHub

Authentication check. Please ignore.
  • Anonymous
  • Sign in
  • Create
  • Ask a question
  • Spaces
  • API Connect
  • Analytic Hybrid Cloud Core
  • Application Performance Management
  • Appsecdev
  • BPM
  • Blockchain
  • Business Transaction Intelligence
  • CAPI
  • CAPI SNAP
  • CICS
  • Cloud Analytics
  • Cloud Automation
  • Cloud Object Storage
  • Cloud marketplace
  • Collaboration
  • Content Services (ECM)
  • Continuous Testing
  • Courses
  • Customer Experience Analytics
  • DB2 LUW
  • DataPower
  • Decision Optimization
  • DevOps Services
  • Developers IBM MX
  • Digital Commerce
  • Digital Experience
  • Finance
  • Global Entrepreneur Program
  • Hadoop
  • Hybrid Cloud Core
  • IBM Cloud platform
  • IBM Design
  • IBM Forms Experience Builder
  • IBM Maximo Developer
  • IBM StoredIQ
  • IBM StoredIQ-Cartridges
  • IIDR
  • ITOA
  • InformationServer
  • Integration Bus
  • Internet of Things
  • Kenexa
  • Linux on Power
  • LinuxONE
  • MDM
  • Mainframe
  • Messaging
  • Node.js
  • ODM
  • Open
  • PowerAI
  • PowerVC
  • Predictive Analytics
  • Product Insights
  • PureData for Analytics
  • Push
  • QRadar App Development
  • Run Book Automation
  • Search Insights
  • Security Core
  • Storage
  • Storage Core
  • Streamsdev
  • Supply Chain Business Network
  • Supply Chain Insights
  • Swift
  • UBX Capture
  • Universal Behavior Exchange
  • UrbanCode
  • WASdev
  • WSRR
  • Watson
  • Watson Campaign Automation
  • Watson Content Hub
  • Watson Marketing Insights
  • dW Answers Help
  • dW Premium
  • developerWorks Sandbox
  • developerWorks Team
  • Watson Health
  • More
  • Tags
  • Questions
  • Users
  • Badges