Spark-MLlib的快速使用之十（ 关联分析）

（1）描述

关联规则挖掘最典型的例子是购物篮分析，通过分析可以知道哪些商品经常被一起购买，从而可以改进商品货架的布局。

（2） 测试数据

r z h k p

z y x w v u t s

s x o n r

x z y m t s q e

z

x z y r q t p

（3） 样例程序

public static void main(String[] args) {

String inputFile;

double minSupport = 0.3;

int numPartition = -1;

inputFile = "sample_fpgrowth.txt";

if (args.length >= 2) {

minSupport = Double.parseDouble(args[1]);

}

if (args.length >= 3) {

numPartition = Integer.parseInt(args[2]);

}

SparkConf sparkConf = new SparkConf().setAppName("JavaFPGrowthExample").setMaster("local");

JavaSparkContext sc = new JavaSparkContext(sparkConf);

JavaRDD<ArrayList<String>> transactions = sc.textFile(inputFile).map(

new Function<String, ArrayList<String>>() {

@Override

public ArrayList<String> call(String s) {

return Lists.newArrayList(s.split(" "));

}

}

);

FPGrowthModel<String> model = new FPGrowth()

.setMinSupport(minSupport)

.setNumPartitions(numPartition)

.run(transactions);

for (FPGrowth.FreqItemset<String> s: model.freqItemsets().toJavaRDD().collect()) {

System.out.println("[" + Joiner.on(",").join(s.javaItems()) + "], " + s.freq());

}

sc.stop();

}

（4）测试结果

[s], 3

[s,x], 3

[s,x,z], 2

[s,z], 2

[r], 3

[r,x], 2

[r,z], 2

[y], 3

[y,s], 2

[y,s,x], 2

[y,s,x,z], 2

[y,s,z], 2

[y,x], 3

[y,x,z], 3

[y,t], 3

[y,t,s], 2

[y,t,s,x], 2

[y,t,s,x,z], 2

[y,t,s,z], 2

[y,t,x], 3

[y,t,x,z], 3

[y,t,z], 3

[y,z], 3

[q], 2

[q,y], 2

[q,y,x], 2

[q,y,x,z], 2

[q,y,t], 2

[q,y,t,x], 2

[q,y,t,x,z], 2

[q,y,t,z], 2

[q,y,z], 2

[q,x], 2

[q,x,z], 2

[q,t], 2

[q,t,x], 2

[q,t,x,z], 2

[q,t,z], 2

[q,z], 2

[x], 4

[x,z], 3

[t], 3

[t,s], 2

[t,s,x], 2

[t,s,x,z], 2

[t,s,z], 2

[t,x], 3

[t,x,z], 3

[t,z], 3

[p], 2

[p,r], 2

[p,r,z], 2

[p,z], 2

[z], 5