问题描述

朴素贝叶斯分类问题 高手请进

网上资料讲mahout 贝叶斯的很多，都是讲准备数据，划分测试集训练集，测试训练样本。。。但是，之后呢？有了这个测试很好的模型，怎么对其他的数据分类啊？
比如我通过mahout进行朴素贝叶斯分类，得到训练集，测试也很好，那么有了这个训练集之后怎么对其他数据进行分类呢。说的越详细越好。

解决方案

运行mahout的朴素贝叶斯分类器

1.准备数据
1.1 下载数据集，并解压

wget http://people.csail.mit.edu/jrennie/20Newsgroups/20news-bydate.tar.gz
tar -xf 20news-bydate.tar.gz
#上传到hdfs
hadoop fs -put 20news-bydate-test .
hadoop fs -put 20news-bydate-train .
1.2 转换格式

#转换为序列文件(sequence files)
mahout seqdirectory -i 20news-bydate-train -o 20news-bydate-train-seq
mahout seqdirectory -i 20news-bydate-test -o 20news-bydate-test-seq
#转换为tf-idf向量
mahout seq2sparse -i 20news-bydate-train-seq -o 20news-bydate-train-vector -lnorm -nv -wt tfidf
mahout seq2sparse -i 20news-bydate-test-seq -o 20news-bydate-test-vector -lnorm -nv -wt tfidf
2. 训练朴素贝叶斯模型
mahout trainnb -i 20news-bydate-train-vectors/tfidf-vectors -el -o model -li labelindex -ow
3. 测试朴素贝叶斯模型
mahout testnb -i 20news-bydate-train-vectors/tfidf-vectors -m model -l labelindex -ow -o test-result
4. 查看训练后的结构
mahout seqdumper -i labelindex

Input Path: labelindex
Key class: class org.apache.hadoop.io.Text Value Class: class org.apache.hadoop.io.IntWritable
Key: alt.atheism: Value: 0
Key: comp.graphics: Value: 1
Key: comp.os.ms-windows.misc: Value: 2
Key: comp.sys.ibm.pc.hardware: Value: 3
Key: comp.sys.mac.hardware: Value: 4
Key: comp.windows.x: Value: 5
Key: misc.forsale: Value: 6
Key: rec.autos: Value: 7
Key: rec.motorcycles: Value: 8
Key: rec.sport.baseball: Value: 9
Key: rec.sport.hockey: Value: 10
Key: sci.crypt: Value: 11
Key: sci.electronics: Value: 12
Key: sci.med: Value: 13
Key: sci.space: Value: 14
Key: soc.religion.christian: Value: 15
Key: talk.politics.guns: Value: 16
Key: talk.politics.mideast: Value: 17
Key: talk.politics.misc: Value: 18
Key: talk.religion.misc: Value: 19
Count: 20