flink-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Todd Lisonbee (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (FLINK-3613) Add standard deviation, mean, variance to list of Aggregations
Date Thu, 17 Mar 2016 22:57:33 GMT

    [ https://issues.apache.org/jira/browse/FLINK-3613?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15200562#comment-15200562
] 

Todd Lisonbee commented on FLINK-3613:
--------------------------------------

I didn't find exact overlap (FLINK-2144 was similar except for Windows, FLINK-2379 is for
vectors but isn't using above interface).

---

Implementing this isn't as easy as extending the existing AggregationFunction abstract class.
 AggregationFunction works for Sum, Min, and Max but isn't general enough for other aggregations.

An aggregation should have three types:
1) the value type - the type being aggregated
2) the aggregate type - the intermediate type that carries all needed data for the aggregation
3) the result type - the result of the aggregation

For example, if you are aggregating doubles in different ways:
SUM - value type is double, aggregation type is double, result type is double
COUNT - value type is double, aggregation type is probably long, result type is long
STANDARD_DEVIATION - value type is double, aggregation type would be a complex type (count,
mean, sum of squares differences from current mean, deltas), result type is double

> Add standard deviation, mean, variance to list of Aggregations
> --------------------------------------------------------------
>
>                 Key: FLINK-3613
>                 URL: https://issues.apache.org/jira/browse/FLINK-3613
>             Project: Flink
>          Issue Type: Improvement
>            Reporter: Todd Lisonbee
>            Priority: Minor
>
> Implement standard deviation, mean, variance for org.apache.flink.api.java.aggregation.Aggregations
> Ideally implementation should be single pass and numerically stable.
> References:
> "Scalable and Numerically Stable Descriptive Statistics in SystemML", Tian et al, International
Conference on Data Engineering 2012
> http://dl.acm.org/citation.cfm?id=2310392
> "The Kahan summation algorithm (also known as compensated summation) reduces the numerical
errors that occur when adding a sequence of finite precision floating point numbers. Numerical
errors arise due to truncation and rounding. These errors can lead to numerical instability
when calculating variance."
> https://en.wikipedia.org/wiki/Kahan_summation_algorithm



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message