mahout学习

参考：http://www.360doc.com/content/14/0117/09/1200324_345883534.shtml

Precondition: 启动Hadoop集群

bin/hdfs namenode -format

sbin/start-dfs.sh(启动Namenode,nodenode相关节点)

sbin/start-yarn.sh(启动ResourceManager，nodeManager相关资源)

bin/hdfs dfsadmin -safemode leave(关闭安全模式)

Note：所有bin/mahout下对应的输入文件，输入文件夹均在HDFS文件目录下

In addition，Mahout下处理的文件必须是SequenceFile文件格式的，故需将txt格式文件转化为SequenceFile文件，如下：

kelvin@Master:~/UntarFile/mahout-0.7-cdh4.5.0$ bin/mahout seqdirectory -input（输入） /TagOutput/txtFile.txt -output（输出）/TagOutput/seqFile.txt --charset UTF-8

相关实践过程如下：

(将sequenceFile文件转化为可读的txt文件)

kelvin@Master:~/UntarFile/mahout-0.7-cdh4.5.0$ bin/mahout seqdumper -i（输入） /TagOutput/clusteredPoints/part-m-00000 -o（输出） ./TagOutput/clusterPoints.txt
Running on hadoop, using /home/kelvin/UntarFile/hadoop2CDH4//bin/hadoop and HADOOP_CONF_DIR=
MAHOUT-JOB: /home/kelvin/UntarFile/mahout-0.7-cdh4.5.0/mahout-examples-0.7-cdh4.5.0-job.jar
14/06/06 02:18:07 INFO common.AbstractJob: Command line arguments: {--endPhase=[2147483647], --input=[/TagOutput/clusteredPoints/part-m-00000], --output=[./TagOutput/clusterPoints.txt], --startPhase=[0], --tempDir=[temp]}
14/06/06 02:18:08 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
14/06/06 02:18:08 INFO driver.MahoutDriver: Program took 1162 ms (Minutes: 0.019366666666666667)
kelvin@Master:~/UntarFile/mahout-0.7-cdh4.5.0$ cd TagOutput/
kelvin@Master:~/UntarFile/mahout-0.7-cdh4.5.0/TagOutput$ ll
total 12
drwxrwxr-x 2 kelvin kelvin 4096 6? 6 02:12 ./
drwxr-xr-x 16 kelvin kelvin 4096 6? 6 02:15 ../
-rw-rw-r-- 1 kelvin kelvin 2767 6? 6 02:18 clusterPoints.txt
kelvin@Master:~/UntarFile/mahout-0.7-cdh4.5.0/TagOutput$ cat clusterPoints.txt
Input Path: /TagOutput/clusteredPoints/part-m-00000
Key class: class org.apache.hadoop.io.IntWritable Value Class: class org.apache.mahout.clustering.classify.WeightedPropertyVectorWritable
Key: 14: Value: wt: 1.0 distance: 13.357142857142858 vec: 2012000317 = [24.000]
Key: 14: Value: wt: 1.0 distance: 6.642857142857143 vec: 2012000318 = [4.000]
Key: 25: Value: wt: 1.0 distance: 11.615384615384592 vec: 2012000319 = [56.000]
Key: 14: Value: wt: 1.0 distance: 7.642857142857142 vec: 2012000320 = [3.000]
Key: 14: Value: wt: 1.0 distance: 11.357142857142858 vec: 2012000321 = [22.000]
Key: 25: Value: wt: 1.0 distance: 0.6153846153846132 vec: 2012000322 = [45.000]
Key: 14: Value: wt: 1.0 distance: 13.357142857142858 vec: 2012000323 = [24.000]
Key: 14: Value: wt: 1.0 distance: 6.642857142857143 vec: 2012000324 = [4.000]
Key: 25: Value: wt: 1.0 distance: 11.615384615384592 vec: 2012000325 = [56.000]
Key: 14: Value: wt: 1.0 distance: 7.642857142857142 vec: 2012000326 = [3.000]
Key: 14: Value: wt: 1.0 distance: 11.357142857142858 vec: 2012000327 = [22.000]
Key: 25: Value: wt: 1.0 distance: 2.384615384615403 vec: 2012000328 = [42.000]
Key: 25: Value: wt: 1.0 distance: 4.61538461538464 vec: 2012000329 = [49.000]
Key: 25: Value: wt: 1.0 distance: 3.384615384615356 vec: 2012000330 = [41.000]
Key: 14: Value: wt: 1.0 distance: 5.642857142857143 vec: 2012000331 = [5.000]
Key: 14: Value: wt: 1.0 distance: 7.642857142857142 vec: 2012000332 = [3.000]
Key: 14: Value: wt: 1.0 distance: 10.357142857142858 vec: 2012000333 = [21.000]
Key: 25: Value: wt: 1.0 distance: 10.38461538461539 vec: 2012000334 = [34.000]
Key: 25: Value: wt: 1.0 distance: 15.384615384615387 vec: 2012000335 = [29.000]
Key: 25: Value: wt: 1.0 distance: 1.3846153846153868 vec: 2012000336 = [43.000]
Key: 25: Value: wt: 1.0 distance: 9.615384615384606 vec: 2012000337 = [54.000]
Key: 25: Value: wt: 1.0 distance: 8.384615384615397 vec: 2012000338 = [36.000]
Key: 14: Value: wt: 1.0 distance: 8.642857142857142 vec: 2012000339 = [2.000]
Key: 14: Value: wt: 1.0 distance: 5.642857142857143 vec: 2012000340 = [5.000]
Key: 20: Value: wt: 1.0 distance: 1.7999999999999972 vec: 2012000341 = [78.000]
Key: 25: Value: wt: 1.0 distance: 9.615384615384606 vec: 2012000342 = [54.000]
Key: 20: Value: wt: 1.0 distance: 13.800000000000018 vec: 2012000343 = [66.000]
Key: 14: Value: wt: 1.0 distance: 3.6428571428571423 vec: 2012000344 = [7.000]
Key: 20: Value: wt: 1.0 distance: 29.200000000000053 vec: 2012000345 = [109.000]
Key: 20: Value: wt: 1.0 distance: 11.800000000000068 vec: 2012000346 = [68.000]
Key: 20: Value: wt: 1.0 distance: 1.7999999999999972 vec: 2012000347 = [78.000]
Key: 25: Value: wt: 1.0 distance: 6.384615384615375 vec: 2012000348 = [38.000]
Count: 32
（将相应的数据结点聚类后输出）

kelvin@Master:~/UntarFile/mahout-0.7-cdh4.5.0$ bin/mahout clusterdump --input /TagOutput/*final(:目录，非文件，最后一次迭代的clusters) --pointsDir /TagOutput/clusteredPoints（最后一次聚类后的点） --output ./TagOutput/clusterResult.txt
Running on hadoop, using /home/kelvin/UntarFile/hadoop2CDH4//bin/hadoop and HADOOP_CONF_DIR=
MAHOUT-JOB: /home/kelvin/UntarFile/mahout-0.7-cdh4.5.0/mahout-examples-0.7-cdh4.5.0-job.jar
14/06/06 02:50:11 INFO common.AbstractJob: Command line arguments: {--dictionaryType=[text], --distanceMeasure=[org.apache.mahout.common.distance.SquaredEuclideanDistanceMeasure], --endPhase=[2147483647], --input=[/TagOutput/*final], --output=[./TagOutput/clusterResult.txt], --outputFormat=[TEXT], --pointsDir=[/TagOutput/clusteredPoints], --startPhase=[0], --tempDir=[temp]}
14/06/06 02:50:12 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
14/06/06 02:50:14 INFO clustering.ClusterDumper: Wrote 3 clusters
14/06/06 02:50:14 INFO driver.MahoutDriver: Program took 2809 ms (Minutes: 0.046816666666666666)
kelvin@Master:~/UntarFile/mahout-0.7-cdh4.5.0$ cd TagOutput/
kelvin@Master:~/UntarFile/mahout-0.7-cdh4.5.0/TagOutput$ ls
clusterPoints.txt clusterResult.txt
kelvin@Master:~/UntarFile/mahout-0.7-cdh4.5.0/TagOutput$ ll
total 16
drwxrwxr-x 2 kelvin kelvin 4096 6? 6 02:30 ./
drwxr-xr-x 16 kelvin kelvin 4096 6? 6 02:15 ../
-rw-rw-r-- 1 kelvin kelvin 2767 6? 6 02:18 clusterPoints.txt
-rw-rw-r-- 1 kelvin kelvin 2108 6? 6 02:50 clusterResult.txt
kelvin@Master:~/UntarFile/mahout-0.7-cdh4.5.0/TagOutput$ cat clusterResult.txt
VL-14{n=14 c=[10.643] r=[9.013]}
        Weight : [props - optional]: Point:
        1.0 : [distance=13.357142857142858]: 2012000317 = [24.000]
        1.0 : [distance=6.642857142857143]: 2012000318 = [4.000]
        1.0 : [distance=7.642857142857142]: 2012000320 = [3.000]
        1.0 : [distance=11.357142857142858]: 2012000321 = [22.000]
        1.0 : [distance=13.357142857142858]: 2012000323 = [24.000]
        1.0 : [distance=6.642857142857143]: 2012000324 = [4.000]
        1.0 : [distance=7.642857142857142]: 2012000326 = [3.000]
        1.0 : [distance=11.357142857142858]: 2012000327 = [22.000]
        1.0 : [distance=5.642857142857143]: 2012000331 = [5.000]
        1.0 : [distance=7.642857142857142]: 2012000332 = [3.000]
        1.0 : [distance=10.357142857142858]: 2012000333 = [21.000]
        1.0 : [distance=8.642857142857142]: 2012000339 = [2.000]
        1.0 : [distance=5.642857142857143]: 2012000340 = [5.000]
        1.0 : [distance=3.6428571428571423]: 2012000344 = [7.000]
VL-20{n=5 c=[79.800] r=[15.419]}
        Weight : [props - optional]: Point:
        1.0 : [distance=1.7999999999999972]: 2012000341 = [78.000]
        1.0 : [distance=13.800000000000018]: 2012000343 = [66.000]
        1.0 : [distance=29.200000000000053]: 2012000345 = [109.000]
        1.0 : [distance=11.800000000000068]: 2012000346 = [68.000]
        1.0 : [distance=1.7999999999999972]: 2012000347 = [78.000]
VL-25{n=13 c=[44.385] r=[8.553]}
        Weight : [props - optional]: Point:
        1.0 : [distance=11.615384615384592]: 2012000319 = [56.000]
        1.0 : [distance=0.6153846153846132]: 2012000322 = [45.000]
        1.0 : [distance=11.615384615384592]: 2012000325 = [56.000]
        1.0 : [distance=2.384615384615403]: 2012000328 = [42.000]
        1.0 : [distance=4.61538461538464]: 2012000329 = [49.000]
        1.0 : [distance=3.384615384615356]: 2012000330 = [41.000]
        1.0 : [distance=10.38461538461539]: 2012000334 = [34.000]
        1.0 : [distance=15.384615384615387]: 2012000335 = [29.000]
        1.0 : [distance=1.3846153846153868]: 2012000336 = [43.000]
        1.0 : [distance=9.615384615384606]: 2012000337 = [54.000]
        1.0 : [distance=8.384615384615397]: 2012000338 = [36.000]
        1.0 : [distance=9.615384615384606]: 2012000342 = [54.000]
        1.0 : [distance=6.384615384615375]: 2012000348 = [38.000]
上述结果的输出等同于在Eclipse中编写的ClusterDumper，如下：

public static void run(Configuration conf, Path input, Path output, DistanceMeasure measure, double t1, double t2,

      double convergenceDelta, int maxIterations) throws Exception {

    Path directoryContainingConvertedInput = new Path(output, DIRECTORY_CONTAINING_CONVERTED_INPUT);

    log.info("Preparing Input");

    TagInputDriver.runJob(input, directoryContainingConvertedInput, "org.apache.mahout.math.RandomAccessSparseVector");

    log.info("Running Canopy to get initial clusters");

    Path canopyOutput = new Path(output, "canopies");

    CanopyDriver.run(new Configuration(), directoryContainingConvertedInput, canopyOutput, measure, t1, t2, false, 0.0,

        false);

    log.info("Running KMeans");

    TagKMeansDriver.run(conf, directoryContainingConvertedInput, new Path(canopyOutput, Cluster.INITIAL_CLUSTERS_DIR

        + "-final"), output, convergenceDelta, maxIterations, true, 0.0, false);

    // run ClusterDumper

    ClusterDumper clusterDumper = new ClusterDumper(new Path(output, "clusters-*-final"), new Path(output,

        "clusteredPoints"));

    clusterDumper.printClusters(null);

  }

其他前提操作：

kelvin@Master:~/UntarFile/hadoop2CDH4$ bin/hadoop fs -put ./../mahout-0.7-cdh4.5.0/output/* /TagOutput
14/06/06 01:53:16 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
kelvin@Master:~/UntarFile/hadoop2CDH4$ bin/hadoop fs -ls /TagOutput
14/06/06 01:53:33 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
Found 15 items
-rw-r--r--   1 kelvin supergroup        194 2014-06-06 01:53 /TagOutput/_policy
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusteredPoints
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-0
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-1
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-10-final
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-2
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-3
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-4
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-5
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-6
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-7
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-8
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/clusters-9
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/data
drwxr-xr-x   - kelvin supergroup          0 2014-06-06 01:53 /TagOutput/random-seeds

mahout学习的更多相关文章

转】Mahout学习路线图
原博文出自于: http://blog.fens.me/hadoop-mahout-roadmap/ 感谢! Mahout学习路线图 Hadoop家族系列文章,主要介绍Hadoop家族产品,常用的项目 ...
mahout第一篇-----Mahout学习路线图
Mahout学习路线图前言 Mahout是Hadoop家族中与众不同的一个成员,是基于一个Hadoop的机器学习和数据挖掘的分布式计算框架.Mahout是一个跨学科产品,同时也是我认为Hadoop家 ...
Mahout学习路线图
转自:http://blog.fens.me/hadoop-mahout-roadmap/ Mahout学习路线图 Hadoop家族系列文章,主要介绍Hadoop家族产品,常用的项目包括Hadoop, ...
Mahout学习路线图-张丹老师
前言 Mahout是Hadoop家族中与众不同的一个成员,是基于一个Hadoop的机器学习和数据挖掘的分布式计算框架.Mahout是一个跨学科产品,同时也是我认为Hadoop家族中,最有竞争力,最难掌 ...
Hadoop里的数据挖掘应用-Mahout——学习笔记<三>
之前有幸在MOOC学院抽中小象学院hadoop体验课. 这是小象学院hadoop2.X的笔记由于平时对数据挖掘做的比较多,所以优先看Mahout方向视频. Mahout有很好的扩展性与容错性(基于H ...
Mahout学习之Mahout简介、安装、配置、入门程序测试
一.Mahout简介查了Mahout的中文意思——驭象的人,再看看Mahout的logo,好吧,想和小黄象happy地玩耍,得顺便陪陪这位驭象人耍耍了... 附logo: (就是他,骑在象头上的那个 ...
Mahout学习之Mahout简单介绍、安装、配置、入门程序測试
一.Mahout简单介绍查了Mahout的中文意思--驭象的人,再看看Mahout的logo,好吧,想和小黄象happy地玩耍,得顺便陪陪这位驭象人耍耍了... 附logo: (就是他,骑在象头上的 ...
mahout学习-1
一. 安装软件需要安装如下文件: java, Eclipse, Maven,Hadoop,mahout 二. 推荐系统简介每天,我们都会对一些事物表达自己的看法,喜欢,或不喜欢,或不在乎.这些都在 ...
Mahout学习资料
Apache Mahout 简介 http://www.ibm.com/developerworks/cn/java/j-mahout/ 从源代码剖析Mahout推荐引擎 http://blog.fe ...

随机推荐

CentOS里vim基本操作
1.关于退出 :wq! ----强制保存退出 :wq ---- 保存退出 :x ----- 作用和:wq 一样 ZZ ---- 作用和:wq一样,(注意Z是大写的,并且不是在命令模式) :q ...
数据库面试sql
问题一:.有三张表,学生表S,课程表C,学生课程表SC 01:写出建表语句答: create table s(id integer primary key,name varchar(20)); cr ...
centOS下NFS服务器的安装配置详解
一.NFS简介 NFS就是Network FileSystem的缩写,最早之前是由Sun公司所发展出来的.他最大的功能就是可以透过网络,让不同的机器.不同的操作系统可以彼此分享个别档案(share f ...
Linux中的LVM和软RAID
在实际工作中,会经常碰到所给的服务器硬盘容量太小,而实际的应用软件中却需要一个容量较大的分区进行数据存储等,除了通过硬件RAID卡来实现合并多硬盘外,其实我们也可以通过软件的方式来实现. 实验 ...
【图论】信手拈来的Prim，Kruskal和Dijkstra
关于三个简单的图论算法 prim,dijkstra和kruskal三个图论的算法,初学者容易将他们搞混,所以放在一起了. prim和kruskal是最小生成树(MST)的算法,dijkstra是单源最 ...
VLC简介及使用说明
一.简介 VLC的全名是Video Lan Client,是一个开源的.跨平台的视频播放器.VLC支持大量的音视频传输.封装和编码格式,完整的功能特性列表可以在这里获得http://www.vi ...
译：微软发布.NET应用架构指南草案
原文<Microsoft Announces Draft .NET Architecture Guidance> 译注:上周微软发布了全新的<.NET应用架构指南>草案,以征求 ...
设计模式总结（《Head First设计模式》学习总结）
写在前面: 学习过程中不仅要熟练掌握技能,理论的消化吸收也必不可少.虽然个人更倾向于学习技术类的东西(短时间的精力投入很快就能看到成效...),但看了很多前辈的经验总结后才知道理论性的东西是绝对不能忽 ...
使用MEF与Castle实现AOP
MEF是微软的一个ioc框架,使用非常方便,我们只需要在需要导出的类上标记[Export],在需要使用的地方[import]就可以使用了.现在我们扩展MEF,在其装配生成实例时,使用Castle Dy ...
Effective Java（1）-创建和销毁对象
Effective Java(1)-创建和销毁对象

mahout学习

mahout学习的更多相关文章

随机推荐

热门专题