mahout-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Weiqing Jin <wq...@yahoo.com.INVALID>
Subject Mahout item based recommender help documentation
Date Thu, 03 Dec 2015 00:22:54 GMT
Hi, I am new to Mahout. I am using Mahout on Cloudera CDH5.3. I believe it has version 0.9.Wondering
how can I get help documentation. Specifically I am trying to use item based recommender algorithm(as
below). I downloaded the Mahout 0.9 distribution files, but not able to find help specifically on
below function, such as what does parameter mean for --numRecommendations etc.Am I missing
some step here? Thanks. 
<!--#yiv6600406099 _filtered #yiv6600406099 {font-family:Calibri;panose-1:2 15 5 2 2 2
4 3 2 4;}#yiv6600406099 #yiv6600406099 p.yiv6600406099MsoNormal, #yiv6600406099 li.yiv6600406099MsoNormal,
#yiv6600406099 div.yiv6600406099MsoNormal {margin:0in;margin-bottom:.0001pt;font-size:11.0pt;font-family:"Calibri",
"sans-serif";}#yiv6600406099 a:link, #yiv6600406099 span.yiv6600406099MsoHyperlink {color:blue;text-decoration:underline;}#yiv6600406099
a:visited, #yiv6600406099 span.yiv6600406099MsoHyperlinkFollowed {color:purple;text-decoration:underline;}#yiv6600406099
span.yiv6600406099EmailStyle17 {font-family:"Calibri", "sans-serif";color:windowtext;}#yiv6600406099
.yiv6600406099MsoChpDefault {font-family:"Calibri", "sans-serif";} _filtered #yiv6600406099
{margin:1.0in 1.0in 1.0in 1.0in;}#yiv6600406099 div.yiv6600406099WordSection1 {}-->mahout
recommenditembased [--input <input> --output <output> --numRecommendations <numRecommendations>  
--usersFile <usersFile> --itemsFile <itemsFile> --filterFile <filterFile>      
--booleanData <booleanData> --maxPrefsPerUser <maxPrefsPerUser>                
--minPrefsPerUser <minPrefsPerUser> --maxSimilaritiesPerItem                   
<maxSimilaritiesPerItem> --maxPrefsInItemSimilarity <maxPrefsInItemSimilarity> 
--similarityClassname <similarityClassname> --threshold <threshold>            
--outputPathForSimilarityMatrix <outputPathForSimilarityMatrix> --randomSeed    <randomSeed>
--sequencefileOutput --help --tempDir <tempDir> --startPhase       <startPhase>
--endPhase <endPhase>]                                            
--similarityClassname (-s) similarityClassname    Name of distributed          
                                                  similarity
measures class to                                                    instantiate,
alternatively                                                      use
one of the predefined                                                       similarities                 
                                                  ([SIMILARITY_COOCCURRENCE,   
                                                  SIMILARITY_LOGLIKELIHOOD,    
                                                  SIMILARITY_TANIMOTO_COEFFICIEN
                                                 
T, SIMILARITY_CITY_BLOCK,                                                       SIMILARITY_COSINE,           
                                                  SIMILARITY_PEARSON_CORRELATION
                                                 
,                                                                               SIMILARITY_EUCLIDEAN_DISTANCE]
                                                 
)       Eric Jin Retail Service Decision Management Eric.Jin@citi.com 224-222-2590   


  
Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message