Java 类名:com.alibaba.alink.pipeline.feature.PCA
Python 类名:PCA
功能介绍
主成分分析,是考察多个变量间相关性一种多元统计方法,研究如何通过少数几个主成分来揭示多个变量间的内部结构,即从原始变量中导出少数几个主成分,使它们尽可能多地保留原始变量的信息,且彼此间互不相关,作为新的综合指标。详细介绍请见维基百科链接wiki。
参数说明
名称 | 中文名称 | 描述 | 类型 | 是否必须? | 取值范围 | 默认值 |
---|---|---|---|---|---|---|
k | 降维后的维度 | 降维后的维度 | Integer | ✓ | [1, +inf) | |
predictionCol | 预测结果列名 | 预测结果列名 | String | ✓ | ||
calculationType | 计算类型 | 计算类型,包含”CORR”, “COV”两种。 | String | “CORR”, “COV” | “CORR” | |
modelFilePath | 模型的文件路径 | 模型的文件路径 | String | null | ||
overwriteSink | 是否覆写已有数据 | 是否覆写已有数据 | Boolean | false | ||
reservedCols | 算法保留列名 | 算法保留列 | String[] | null | ||
selectedCols | 选中的列名数组 | 计算列对应的列名列表 | String[] | null | ||
vectorCol | 向量列名 | 向量列对应的列名,默认值是null | String | null | ||
numThreads | 组件多线程线程个数 | 组件多线程线程个数 | Integer | 1 | ||
modelStreamFilePath | 模型流的文件路径 | 模型流的文件路径 | String | null | ||
modelStreamScanInterval | 扫描模型路径的时间间隔 | 描模型路径的时间间隔,单位秒 | Integer | 10 | ||
modelStreamStartTime | 模型流的起始时间 | 模型流的起始时间。默认从当前时刻开始读。使用yyyy-mm-dd hh:mm:ss.fffffffff格式,详见Timestamp.valueOf(String s) | String | null |
代码示例
Python 代码
from pyalink.alink import *
import pandas as pd
useLocalEnv(1)
df = pd.DataFrame([
[0.0,0.0,0.0],
[0.1,0.2,0.1],
[0.2,0.2,0.8],
[9.0,9.5,9.7],
[9.1,9.1,9.6],
[9.2,9.3,9.9]
])
# batch source
inOp = BatchOperator.fromDataframe(df, schemaStr='x1 double, x2 double, x3 double')
pca = PCA().setK(2).setSelectedCols(["x1","x2","x3"]).setPredictionCol("pred")
# train
model = pca.fit(inOp)
# batch predict
model.transform(inOp).print()
# stream predict
inStreamOp = StreamOperator.fromDataframe(df, schemaStr='x1 double, x2 double, x3 double')
model.transform(inStreamOp).print()
StreamOperator.execute()
Java 代码
package javatest.com.alibaba.alink.pipeline.feature;
import org.apache.flink.types.Row;
import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
import com.alibaba.alink.operator.stream.StreamOperator;
import com.alibaba.alink.operator.stream.source.MemSourceStreamOp;
import com.alibaba.alink.pipeline.feature.PCA;
import com.alibaba.alink.pipeline.feature.PCAModel;
import org.junit.Test;
import java.util.Arrays;
import java.util.List;
public class PcaTest {
@Test
public void testPca() throws Exception {
List <Row> df = Arrays.asList(
Row.of(0.0, 0.0, 0.0),
Row.of(0.1, 0.2, 0.1),
Row.of(0.2, 0.2, 0.8),
Row.of(9.0, 9.5, 9.7),
Row.of(9.1, 9.1, 9.6),
Row.of(9.2, 9.3, 9.9)
);
BatchOperator <?> inOp = new MemSourceBatchOp(df, "x1 double, x2 double, x3 double");
MemSourceStreamOp inStreamOp = new MemSourceStreamOp(df, "x1 double, x2 double, x3 double");
PCA pca = new PCA()
.setK(2)
.setSelectedCols(new String[] {"x1", "x2", "x3"}).setPredictionCol("pred");
PCAModel model = pca.fit(inOp);
model.transform(inOp).print();
model.transform(inStreamOp).print();
StreamOperator.execute();
}
}
结果
| x1 | x2 | x3 | pred | | —- | —- | —- | —- |
| 0.0000 | 0.0000 | 0.0000 | -1.6404909810453345 -0.0251812826908675 |
| 0.1000 | 0.2000 | 0.1000 | -1.5946357760302712 -0.037364200387782764 |
| 0.2000 | 0.2000 | 0.8000 | -1.5048402139720405 0.06485201225195414 |
| 9.0000 | 9.5000 | 9.7000 | 1.587547449494739 -0.02506612043660217 |
| 9.1000 | 9.1000 | 9.6000 | 1.5421273389336387 0.0022882493013524074 |
| 9.2000 | 9.3000 | 9.9000 | 1.6102921826192689 0.020471341961945777 |