UPDATED ACTUAL Questions and
CORRECT Answers
Synchronization - CORRECT ANSWER - The act of telling the analytics server about
updates to documents coded on the Review Field.
Categorization last completed on - CORRECT ANSWER - Doesn't necessarily mean the
model rebuilt. It generally just means that a heartbeat was conducted. Of course, if a model build
is in progress, you'll see that in the 'Categorization Status' field.
Active learning projects automatically rebuild the model and rerank the documents
approximately every 20 minutes. - CORRECT ANSWER - True
A new Active Learning project will automatically create its own Analytics Categorization Set. -
CORRECT ANSWER - True. Unlike categorization within conceptual analytics, the
Active Learning Categorization set can't be edited.
What fields should you use for the view in the Analytics Categorization Set tab? - CORRECT
ANSWER - Name
Categorization Status
Categorization Last Completed On
Categorization Last Run Error
Synchronization Status
Synchronization Last Completed On
Synchronization Last Run Error
Update Ranks Status
Update Ranks Last Completed On
Update Ranks Last Run Error
, Does conceptual categorization always require a saved search to define the documents to be
categorized? - CORRECT ANSWER - Yes
Document types to exclude from Active Learning - CORRECT ANSWER - - Empty
documents (i.e. documents containing 0 characters)
- Documents containing only numbers, punctuation, and whitespace (i.e. they become empty
after filtering)
- Emails with no subject and no body apart from URLs, numbers, or stop words (empty after
filtering)
- Media files containing placeholder text such as "jpeg image", "windows movie", or "mpeg
1280x1080"
- Excel spreadsheets whose responsiveness is determined by the formatting (e.g. responsive only
if the entry in column C, row 12 is set to "false").
- Documents whose text contains only placeholder language such as "failed to extract text" or
"no image present"
Elusion Test - CORRECT ANSWER - An Active Learning QC approach to measure the
performance of the model.
The saved search for Active Learning should only return the extracted text field. - CORRECT
ANSWER - True
The account name, WorldCom, occurs too many times in the resulting cluster titles. What should
you do to resolve the problem with the client name, WorldCom? - CORRECT ANSWER -
Add the client name to the Concept Stop Word list.
There are 25 scanned, handwritten documents that have been OCRed in an effort to make them
searchable. What acceptable actions can you take to properly handle the OCRed documents? -
CORRECT ANSWER - - exclude the OCRed documents and handle them manually
- include the OCRed documents in the searchable set only