[SOLR-10590] Add Cross Correlation Stream Evaluator

• Type: New Feature
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Fix Version/s: 7.0
• Component/s: streaming expressions
• Assignee: Unassigned
• Reporter: Joel Bernstein
• Created: 01/May/17 01:38
• Updated: 08/Jun/19 15:21

Description

Now that we have basic correlation (SOLR-10582) implemented we can use it at as a basis for cross correlation. Apache commons math apparently does not have a cross correlation implementation, so it will need to be implemented.

The basic approach taken will be to slide columnA along columnB and perform the correlation calculation until correlation is either 1.0 or max lag has been reached.

Both auto-correlation and auto-regression can be built on top the cross-collation function.

If anyone has an alternative more efficient approach or knows of an existing implementation that can be plugged in, please let me know.

Syntax:

let(a=expr, 
    b=expr, 
    c=col(a, fieldName), 
    d=col(b, fieldName), 
    tuple(corr=xcorr(c,d,10)))

Issue Links

• is related to: SOLR-10651 Streaming Expressions statistical functions library