How do you map the output of RowSimilarity to documents? What I really
need is to create an association of

    doc1 --> docn, docm, doci, etc.

The output of rowsimilarity looks like

    rowid --> vector of rowids : distances

for example:

    Key: 0: Value: {14458:0.2966480826934176,11399:0.30290014772966095,
    12793:0.22009858979452146,3275:0.1871791030103281,
    14613:0.3534278632679437,4411:0.2516380602790199,
    17520:0.3139731583634198,13611:0.18968888212315968,
    14354:0.17673965754661425,0:1.0000000000000004}

It would be nice to use the same keys as they are output by seq2aparse,
in my case named vectors so file names would appear in the output as
rowids. Creating my association would be trivial.

Have I missed a dictionary containing rowid to docid(name) mapping?

Docindex is ur answer

Sent from my iPhone

On Mar 20, 2012, at 12:28 PM, Pat Ferrel <[EMAIL PROTECTED]> wrote:

> How do you map the output of RowSimilarity to documents? What I really need is to create an association of
>
>   doc1 --> docn, docm, doci, etc.
>
> The output of rowsimilarity looks like
>
>   rowid --> vector of rowids : distances
>
> for example:
>
>   Key: 0: Value: {14458:0.2966480826934176,11399:0.30290014772966095,
>   12793:0.22009858979452146,3275:0.1871791030103281,
>   14613:0.3534278632679437,4411:0.2516380602790199,
>   17520:0.3139731583634198,13611:0.18968888212315968,
>   14354:0.17673965754661425,0:1.0000000000000004}
>
> It would be nice to use the same keys as they are output by seq2aparse, in my case named vectors so file names would appear in the output as rowids. Creating my association would be trivial.
>
> Have I missed a dictionary containing rowid to docid(name) mapping?
>

I should have been more elaborate in my previous reply.
RowId job creates a matrix which is of type <IntWritable, VectorWritable> and a docIndex <IntWritable, Text>

docIndex is a map of the rowId to the keys generated from seq2sparse.

What you would need to do is to join the output of RowSimilarity to docIndex to get the format u r looking for.
Hope that helps.
Suneel
________________________________
 From: Suneel Marthi <[EMAIL PROTECTED]>
To: "[EMAIL PROTECTED]" <[EMAIL PROTECTED]>
Sent: Tuesday, March 20, 2012 1:41 PM
Subject: Re: RowSimilarityJob
 
Docindex is ur answer

Sent from my iPhone

On Mar 20, 2012, at 12:28 PM, Pat Ferrel <[EMAIL PROTECTED]> wrote:

> How do you map the output of RowSimilarity to documents? What I really need is to create an association of
>
>   doc1 --> docn, docm, doci, etc.
>
> The output of rowsimilarity looks like
>
>   rowid --> vector of rowids : distances
>
> for example:
>
>   Key: 0: Value: {14458:0.2966480826934176,11399:0.30290014772966095,
>   12793:0.22009858979452146,3275:0.1871791030103281,
>   14613:0.3534278632679437,4411:0.2516380602790199,
>   17520:0.3139731583634198,13611:0.18968888212315968,
>   14354:0.17673965754661425,0:1.0000000000000004}
>
> It would be nice to use the same keys as they are output by seq2aparse, in my case named vectors so file names would appear in the output as rowids. Creating my association would be trivial.
>
> Have I missed a dictionary containing rowid to docid(name) mapping?
>

What is the value created to describe simlarity by RowSimilarityJob? The
paper which describes how the algorithm is implemented doesn't describe
the various similarity values returned by mahout. It seems to focus on
cooccurrences.

For SIMILARITY_COSINE is the value = cosine or 1 - cosine?

Is the value calculated after cooccurrences determines similar docs
independently?

The code is very difficult to read so a little help would be appreciated.