[
https://issues.apache.org/jira/browse/MAHOUT-1464?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14015788#comment-14015788
]
Pat Ferrel commented on MAHOUT-1464:
------------------------------------
Looks like DrmLike may have been refactored since this patch was written.
[~dlyubimov] The following patch code has an error at "elem" saying "Missing
parameter type 'elem'" Looking at the scaladocs I tracked back to the DrmLike
trait and see no way to .mapBlock on it. Has something been refactored here?
The .nonZeroes() is a java sparse vector iterator I think. This worked about a
month ago so thought you might have an idea how things have changed?
{code:scala}
def computeIndicators(drmBtA: DrmLike[Int], numUsers: Int,
maxInterestingItemsPerThing: Int,
bcastNumInteractionsB: Broadcast[Vector],
bcastNumInteractionsA: Broadcast[Vector],
crossCooccurrence: Boolean = true) = {
drmBtA.mapBlock() {
case (keys, block) =>
val llrBlock = block.like()
val numInteractionsB: Vector = bcastNumInteractionsB
val numInteractionsA: Vector = bcastNumInteractionsA
for (index <- 0 until keys.size) {
val thingB = keys(index)
// PriorityQueue to select the top-k items
val topItemsPerThing = new
mutable.PriorityQueue[(Int,Double)]()(orderByScore)
block(index, ::).nonZeroes().foreach { elem => //!!!!!!!!!!!!! Error:
"Missing parameter type 'elem'"
val thingA = elem.index
val cooccurrences = elem.get
{code}
> Cooccurrence Analysis on Spark
> ------------------------------
>
> Key: MAHOUT-1464
> URL: https://issues.apache.org/jira/browse/MAHOUT-1464
> Project: Mahout
> Issue Type: Improvement
> Components: Collaborative Filtering
> Environment: hadoop, spark
> Reporter: Pat Ferrel
> Assignee: Sebastian Schelter
> Fix For: 1.0
>
> Attachments: MAHOUT-1464.patch, MAHOUT-1464.patch, MAHOUT-1464.patch,
> MAHOUT-1464.patch, MAHOUT-1464.patch, MAHOUT-1464.patch, run-spark-xrsj.sh
>
>
> Create a version of Cooccurrence Analysis (RowSimilarityJob with LLR) that
> runs on Spark. This should be compatible with Mahout Spark DRM DSL so a DRM
> can be used as input.
> Ideally this would extend to cover MAHOUT-1422. This cross-cooccurrence has
> several applications including cross-action recommendations.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
