Python+Spark2.0+hadoop学习笔记——Python Spark MLlib决策树二分类
作者：互联网
决策树是在数据分类问题中的一种常用且经典的机器学习方法，在本例里使用决策树模型来分析StumbleUpon数据集，来预测网页是暂时的(ephemeral)还是长青的(evergreen)，并且调教参数找出最佳的参数组合，来提高预测准确度。
像这类经典的二分类问题，在python中调包来做的话会非常容易，且分析手段也有很多。但是现在的练习任务是使用Spark来对着类问题进行处理，因此，下面将开始介绍使用Spark进行二分类问题的过程。
第一步：分析数据的特性
我们在本例中要使用的数据来自于Kaggle官网的数据，这份数据的维度很大，有些数据的列是没有意义的，在选择时候要规避。数据虽然说维度很大，但是很多列都涉及到了一些文本的内容，因此在这个问题中不考虑维度灾难，进行初步的筛选之后就进行建模了。
第二步：启动Ipython Notebook
cd ~/pythonwork/ipynotebook
PYSPARK_DRIVER_PYTHON=ipython PYSPARK_DRIVER_PYTHON_OPTS=“notebook" MASTER=local[*] pyspark
第三步：进行数据准备
原始数据需要变换成训练模型所需的数据格式LabeledPoint，并以随机方式按照8:1:1比例把数据分割为3个部分trainData、validationData和testData。
from pyspark.mllib.regression import LabeledPoint
def SetPath(sc):
global Path
if sc.master[0:5]=="local" :
Path="file:/home/jorlinlee/pythonsparkexample/PythonProject/"
else:
Path="hdfs://master:9000/user/jorlinlee/"
def get_mapping(rdd, idx):
return rdd.map(lambda fields: fields[idx]).distinct().zipWithIndex().collectAsMap()
def extract_label(record):
label=(record[-1])
return float(label)
def extract_features(field,categoriesMap,featureEnd):
categoryIdx = categoriesMap[field[3]]
categoryFeatures = np.zeros(len(categoriesMap))
categoryFeatures[categoryIdx] = 1
numericalFeatures=[convert_float(field) for field in field[4: featureEnd]]
return np.concatenate(( categoryFeatures, numericalFeatures))
def convert_float(x):
return (0 if x=="?" else float(x))
def PrepareData(sc):
print("Data loading...")
rawDataWithHeader = sc.textFile(Path+"data/train.tsv")
header = rawDataWithHeader.first()
rawData = rawDataWithHeader.filter(lambda x:x !=header)
rData=rawData.map(lambda x: x.replace("\"", ""))
lines = rData.map(lambda x: x.split("\t"))
print("The number of data" + str(lines.count()))
categoriesMap = lines.map(lambda fields: fields[3]).distinct().zipWithIndex().collectAsMap()
labelpointRDD = lines.map( lambda r:LabeledPoint(
extract_label(r),
extract_features(r,categoriesMap,len(r) - 1)))
(trainData, validationData, testData) = labelpointRDD.randomSplit([8, 1, 1])
print("The number of trainData:" + str(trainData.count()) +
" The number of validationData:" + str(validationData.count()) +
" The number of testData:" + str(testData.count()))
return (trainData, validationData, testData, categoriesMap)
第四步：对模型进行训练
def PredictData(sc,model,categoriesMap):
print("Data loading.....")
rawDataWithHeader = sc.textFile(Path+"data/test.tsv")
header = rawDataWithHeader.first()
rawData = rawDataWithHeader.filter(lambda x:x !=header)
rData=rawData.map(lambda x: x.replace("\"", ""))
lines = rData.map(lambda x: x.split("\t"))
print("The number of data：" + str(lines.count()) )
dataRDD = lines.map(lambda r: ( r[0] ,
extract_features(r,categoriesMap,len(r) )))
DescDict = {
0: "ephemeral",
1: "evergreen"
}
for data in dataRDD.take(10):
predictResult = model.predict(data[1])
print("The web: " +str(data[0])+"\n" +"Predict:"+ str(predictResult)+"Illustration"+DescDict[predictResult] +"\n")
第五步：评估模型的准确率（使用AUC）
def evaluateModel(model, validationData):
score = model.predict(validationData.map(lambda p: p.features))
scoreAndLabels=score.zip(validationData.map(lambda p: p.label))
metrics = BinaryClassificationMetrics(scoreAndLabels)
AUC=metrics.areaUnderROC
return(AUC)
第六步：评估并确定模型的参数（需要确定impurityParm、maxDepthParm和maxBinsParm）
def trainEvaluateModel(trainData,validationData,
impurityParm, maxDepthParm, maxBinsParm):
startTime = time()
model = DecisionTree.trainClassifier(trainData,
numClasses=2, categoricalFeaturesInfo={},
impurity=impurityParm,
maxDepth=maxDepthParm,
maxBins=maxBinsParm)
AUC = evaluateModel(model, validationData)
duration = time() - startTime
print ("Parameter" + \
" impurity="+str(impurityParm) +\
" maxDepth="+str(maxDepthParm) + \
" maxBins="+str(maxBinsParm) +\
" Time="+str(duration) + \
" AUC = " + str(AUC) )
return (AUC,duration, impurityParm, maxDepthParm, maxBinsParm,model)
def evalParameter(trainData, validationData, evalparm,
impurityList, maxDepthList, maxBinsList):
metrics = [trainEvaluateModel(trainData, validationData,
impurity,maxDepth, maxBins )
for impurity in impurityList
for maxDepth in maxDepthList
for maxBins in maxBinsList ]
if evalparm=="impurity":
IndexList=impurityList[:]
elif evalparm=="maxDepth":
IndexList=maxDepthList[:]
elif evalparm=="maxBins":
IndexList=maxBinsList[:]
df = pd.DataFrame(metrics,index=IndexList,
columns=['AUC', 'duration','impurity', 'maxDepth', 'maxBins','model'])
showchart(df,evalparm,'AUC','duration',0.5,0.7 )
使用可视化方法来作为参数选择的参考
def showchart(df,evalparm ,barData,lineData,yMin,yMax):
ax = df[barData].plot(kind='bar', title =evalparm,figsize=(10,6),legend=True, fontsize=12)
ax.set_xlabel(evalparm,fontsize=12)
ax.set_ylim([yMin,yMax])
ax.set_ylabel(barData,fontsize=12)
ax2 = ax.twinx()
ax2.plot(df[[lineData ]].values, linestyle='-', marker='o', linewidth=2.0,color='r')
plt.show()
找出准确率最高的参数组合
def evalAllParameter(trainData, validationData,
impurityList, maxDepthList, maxBinsList):
metrics = [trainEvaluateModel(trainData, validationData,
impurity,maxDepth, maxBins )
for impurity in impurityList
for maxDepth in maxDepthList
for maxBins in maxBinsList ]
Smetrics = sorted(metrics, key=lambda k: k[0], reverse=True)
bestParameter=Smetrics[0]
print("The best parameter：impurity:" + str(bestParameter[2]) +
" ,maxDepth:" + str(bestParameter[3]) +
" ,maxBins:" + str(bestParameter[4]) +
" ,AUC = " + str(bestParameter[0]))
return bestParameter[5]
def parametersEval(trainData, validationData):
print("----- impurity ---------")
evalParameter(trainData, validationData,"impurity",
impurityList=["gini", "entropy"],
maxDepthList=[10],
maxBinsList=[10 ])
给定参数的选择范围
print("----- maxDepth ---------")
evalParameter(trainData, validationData,"maxDepth",
impurityList=["gini"],
maxDepthList=[3, 5, 10, 15, 20, 25],
maxBinsList=[10])
print("----- maxBins ---------")
evalParameter(trainData, validationData,"maxBins",
impurityList=["gini"],
maxDepthList =[10],
maxBinsList=[3, 5, 10, 50, 100, 200 ])
第七步：Spark的相关设置
def SetLogger( sc ):
logger = sc._jvm.org.apache.log4j
logger.LogManager.getLogger("org"). setLevel( logger.Level.ERROR )
logger.LogManager.getLogger("akka").setLevel( logger.Level.ERROR )
logger.LogManager.getRootLogger().setLevel(logger.Level.ERROR)
def SetPath(sc):
global Path
if sc.master[0:5]=="local" :
Path="file:/home/jorlinlee/pythonsparkexample/PythonProject/"
else:
Path="hdfs://master:9000/user/jorlinlee/"
def CreateSparkContext():
conf = SparkConf().setMaster("local[*]").setAppName("DTree")
sc = SparkContext(conf = conf)
print ("master="+sc.master)
SetLogger(sc)
SetPath(sc)
return (sc)
sc.stop()
第八步：运行主程序，得到十个预测结果，并且给出分类规则
if __name__ == "__main__":
print("RunDecisionTreeBinary")
sc=CreateSparkContext()
print("==========Preparing===============")
(trainData, validationData, testData, categoriesMap) =PrepareData(sc)
trainData.persist(); validationData.persist(); testData.persist()
print("==========Evaluating===============")
(AUC,duration, impurityParm, maxDepthParm, maxBinsParm,model)= \
trainEvaluateModel(trainData, validationData, "entropy", 10, 200)
if (len(sys.argv) == 2) and (sys.argv[1]=="-e"):
parametersEval(trainData, validationData)
elif (len(sys.argv) == 2) and (sys.argv[1]=="-a"):
print("-----The best parameter---------")
model=evalAllParameter(trainData, validationData,
["gini", "entropy"],
[3, 5, 10, 15, 20, 25],
[3, 5, 10, 50, 100, 200 ])
print("==========Testing===============")
auc = evaluateModel(model, testData)
print("Test best model, AUC:" + str(auc))
print("==========Predicting===============")
PredictData(sc, model, categoriesMap)
print(model.toDebugString())
结果：
RunDecisionTreeBinary
master=local[*]
==========Preparing===============
Data loading...
The number of data7395
The number of trainData:5925 The number of validationData:725 The number of testData:745
==========Evaluating===============
Parameter impurity=entropy maxDepth=10 maxBins=200 Time=6.611667633056641 AUC = 0.6579354553492485
==========Testing===============
Test best model, AUC:0.6323763091368725
==========Predicting===============
Data loading.....
The number of data：3171
The web: http://www.lynnskitchenadventures.com/2009/04/homemade-enchilada-sauce.html
Predict:1.0Illustrationevergreen
The web: http://lolpics.se/18552-stun-grenade-ar
Predict:0.0Illustrationephemeral
The web: http://www.xcelerationfitness.com/treadmills.html
Predict:0.0Illustrationephemeral
The web: http://www.bloomberg.com/news/2012-02-06/syria-s-assad-deploys-tactics-of-father-to-crush-revolt-threatening-reign.html
Predict:0.0Illustrationephemeral
The web: http://www.wired.com/gadgetlab/2011/12/stem-turns-lemons-and-limes-into-juicy-atomizers/
Predict:0.0Illustrationephemeral
The web: http://www.latimes.com/health/boostershots/la-heb-fat-tax-denmark-20111013,0,2603132.story
Predict:0.0Illustrationephemeral
The web: http://www.howlifeworks.com/a/a?AG_ID=1186&cid=7340ci
Predict:1.0Illustrationevergreen
The web: http://romancingthestoveblog.wordpress.com/2010/01/13/sweet-potato-ravioli-with-lemon-sage-brown-butter-sauce/
Predict:1.0Illustrationevergreen
The web: http://www.funniez.net/Funny-Pictures/turn-men-down.html
Predict:0.0Illustrationephemeral
The web: http://youfellasleepwatchingadvd.com/
Predict:1.0Illustrationevergreen
DecisionTreeModel classifier of depth 10 with 419 nodes
If (feature 31 <= 1458.5)
If (feature 2 <= 0.5)
If (feature 23 <= 0.0391787705)
If (feature 16 <= 0.719848485)
If (feature 1 <= 0.5)
If (feature 14 <= 0.8458745000000001)
If (feature 8 <= 0.5)
If (feature 0 <= 0.5)
If (feature 29 <= 34.5)
If (feature 16 <= 0.5433298440000001)
Predict: 0.0
Else (feature 16 > 0.5433298440000001)
Predict: 1.0
Else (feature 29 > 34.5)
Predict: 0.0
Else (feature 0 > 0.5)
If (feature 33 <= 5.5)
If (feature 34 <= 0.2243290195)
Predict: 1.0
Else (feature 34 > 0.2243290195)
Predict: 0.0
Else (feature 33 > 5.5)
Predict: 0.0
Else (feature 8 > 0.5)
If (feature 14 <= 0.4711965)
If (feature 14 <= 0.46250199999999997)
Predict: 1.0
Else (feature 14 > 0.46250199999999997)
Predict: 0.0
Else (feature 14 > 0.4711965)
Predict: 1.0
Else (feature 14 > 0.8458745000000001)
Predict: 0.0
Else (feature 1 > 0.5)
If (feature 18 <= 9.702515E-4)
If (feature 34 <= 4.495E-4)
If (feature 23 <= 0.0197483205)
If (feature 35 <= 0.043529472)
Predict: 1.0
Else (feature 35 > 0.043529472)
Predict: 0.0
Else (feature 23 > 0.0197483205)
Predict: 1.0
Else (feature 34 > 4.495E-4)
Predict: 0.0
Else (feature 18 > 9.702515E-4)
If (feature 16 <= 0.4610159345)
Predict: 1.0
Else (feature 16 > 0.4610159345)
If (feature 19 <= 0.050450614000000005)
If (feature 17 <= 0.07201618500000001)
Predict: 1.0
Else (feature 17 > 0.07201618500000001)
Predict: 0.0
Else (feature 19 > 0.050450614000000005)
Predict: 1.0
Else (feature 16 > 0.719848485)
If (feature 19 <= 0.10272814250000001)
If (feature 23 <= 0.035596906)
If (feature 18 <= 0.3340675475)
If (feature 31 <= 1350.5)
If (feature 6 <= 0.5)
Predict: 1.0
Else (feature 6 > 0.5)
Predict: 0.0
Else (feature 31 > 1350.5)
If (feature 18 <= 0.0885694875)
Predict: 1.0
Else (feature 18 > 0.0885694875)
Predict: 0.0
Else (feature 18 > 0.3340675475)
Predict: 0.0
Else (feature 23 > 0.035596906)
Predict: 0.0
Else (feature 19 > 0.10272814250000001)
Predict: 1.0
Else (feature 23 > 0.0391787705)
If (feature 26 <= 0.614921223)
If (feature 15 <= 1.882617912)
If (feature 18 <= 0.0618079675)
If (feature 25 <= 0.18241650199999998)
If (feature 14 <= 0.88331)
If (feature 15 <= 1.5707610145)
Predict: 0.0
Else (feature 15 > 1.5707610145)
Predict: 1.0
Else (feature 14 > 0.88331)
Predict: 1.0
Else (feature 25 > 0.18241650199999998)
If (feature 26 <= 0.18868168899999999)
Predict: 0.0
Else (feature 26 > 0.18868168899999999)
If (feature 32 <= 184.5)
If (feature 23 <= 0.047017412999999994)
Predict: 1.0
Else (feature 23 > 0.047017412999999994)
Predict: 0.0
Else (feature 32 > 184.5)
Predict: 0.0
Else (feature 18 > 0.0618079675)
If (feature 15 <= 1.813996156)
If (feature 15 <= 1.424541885)
Predict: 1.0
Else (feature 15 > 1.424541885)
If (feature 23 <= 0.105398118)
If (feature 15 <= 1.7089758339999999)
Predict: 0.0
Else (feature 15 > 1.7089758339999999)
Predict: 1.0
Else (feature 23 > 0.105398118)
Predict: 1.0
Else (feature 15 > 1.813996156)
Predict: 1.0
Else (feature 15 > 1.882617912)
If (feature 14 <= 0.48163900000000004)
If (feature 15 <= 2.718090381)
If (feature 17 <= 0.08549820050000001)
If (feature 4 <= 0.5)
Predict: 0.0
Else (feature 4 > 0.5)
If (feature 23 <= 0.0719649975)
Predict: 1.0
Else (feature 23 > 0.0719649975)
Predict: 0.0
Else (feature 17 > 0.08549820050000001)
If (feature 25 <= 0.2354568495)
Predict: 0.0
Else (feature 25 > 0.2354568495)
If (feature 14 <= 0.37942050000000005)
Predict: 0.0
Else (feature 14 > 0.37942050000000005)
Predict: 1.0
Else (feature 15 > 2.718090381)
If (feature 35 <= 0.3323970035)
If (feature 16 <= 0.641086983)
If (feature 16 <= 0.29731728749999997)
Predict: 1.0
Else (feature 16 > 0.29731728749999997)
Predict: 0.0
Else (feature 16 > 0.641086983)
If (feature 15 <= 3.024357284)
Predict: 0.0
Else (feature 15 > 3.024357284)
Predict: 1.0
Else (feature 35 > 0.3323970035)
Predict: 1.0
Else (feature 14 > 0.48163900000000004)
If (feature 20 <= 0.408191065)
If (feature 33 <= 5.5)
If (feature 20 <= 0.392865142)
If (feature 15 <= 3.219436535)
Predict: 0.0
Else (feature 15 > 3.219436535)
Predict: 1.0
Else (feature 20 > 0.392865142)
Predict: 1.0
Else (feature 33 > 5.5)
Predict: 1.0
Else (feature 20 > 0.408191065)
If (feature 20 <= 16.0)
Predict: 0.0
Else (feature 20 > 16.0)
If (feature 34 <= 0.0059084325)
Predict: 0.0
Else (feature 34 > 0.0059084325)
If (feature 25 <= 0.296182712)
Predict: 1.0
Else (feature 25 > 0.296182712)
Predict: 0.0
Else (feature 26 > 0.614921223)
If (feature 17 <= 0.29483385900000003)
If (feature 26 <= 3.845238095)
If (feature 32 <= 90.5)
If (feature 25 <= 0.260248616)
If (feature 20 <= 0.6091436754999999)
If (feature 14 <= 0.651891)
Predict: 0.0
Else (feature 14 > 0.651891)
Predict: 1.0
Else (feature 20 > 0.6091436754999999)
Predict: 0.0
Else (feature 25 > 0.260248616)
If (feature 20 <= 0.6210849915000001)
If (feature 25 <= 0.263034435)
Predict: 0.0
Else (feature 25 > 0.263034435)
Predict: 1.0
Else (feature 20 > 0.6210849915000001)
Predict: 0.0
Else (feature 32 > 90.5)
If (feature 35 <= 0.068156843)
If (feature 15 <= 1.199737533)
If (feature 1 <= 0.5)
Predict: 0.0
Else (feature 1 > 0.5)
Predict: 1.0
Else (feature 15 > 1.199737533)
Predict: 0.0
Else (feature 35 > 0.068156843)
If (feature 31 <= 1059.5)
Predict: 0.0
Else (feature 31 > 1059.5)
Predict: 1.0
Else (feature 26 > 3.845238095)
Predict: 0.0
Else (feature 17 > 0.29483385900000003)
If (feature 15 <= 2.5117194654999997)
Predict: 1.0
Else (feature 15 > 2.5117194654999997)
Predict: 0.0
Else (feature 2 > 0.5)
If (feature 14 <= 0.5029295)
If (feature 33 <= 7.5)
If (feature 32 <= 20.5)
Predict: 0.0
Else (feature 32 > 20.5)
If (feature 32 <= 66.5)
Predict: 1.0
Else (feature 32 > 66.5)
Predict: 0.0
Else (feature 33 > 7.5)
Predict: 1.0
Else (feature 14 > 0.5029295)
If (feature 35 <= 0.053987811499999996)
If (feature 26 <= 0.092158035)
Predict: 0.0
Else (feature 26 > 0.092158035)
Predict: 1.0
Else (feature 35 > 0.053987811499999996)
If (feature 27 <= 0.5)
Predict: 1.0
Else (feature 27 > 0.5)
If (feature 25 <= 0.21417068)
If (feature 16 <= 0.617740727)
Predict: 0.0
Else (feature 16 > 0.617740727)
Predict: 1.0
Else (feature 25 > 0.21417068)
If (feature 24 <= 0.5)
Predict: 1.0
Else (feature 24 > 0.5)
Predict: 0.0
Else (feature 31 > 1458.5)
If (feature 3 <= 0.5)
If (feature 0 <= 0.5)
If (feature 1 <= 0.5)
If (feature 29 <= 25.5)
If (feature 19 <= 0.078501401)
If (feature 23 <= 0.052976552499999996)
If (feature 33 <= 6.5)
If (feature 26 <= 0.1836998015)
Predict: 1.0
Else (feature 26 > 0.1836998015)
If (feature 33 <= 0.5)
Predict: 0.0
Else (feature 33 > 0.5)
Predict: 1.0
Else (feature 33 > 6.5)
If (feature 8 <= 0.5)
If (feature 16 <= 0.387492143)
Predict: 0.0
Else (feature 16 > 0.387492143)
Predict: 1.0
Else (feature 8 > 0.5)
Predict: 0.0
Else (feature 23 > 0.052976552499999996)
If (feature 9 <= 0.5)
If (feature 16 <= 0.762858822)
Predict: 0.0
Else (feature 16 > 0.762858822)
Predict: 1.0
Else (feature 9 > 0.5)
If (feature 14 <= 0.2909735)
Predict: 1.0
Else (feature 14 > 0.2909735)
Predict: 0.0
Else (feature 19 > 0.078501401)
If (feature 15 <= 2.5904545455)
If (feature 32 <= 51.5)
If (feature 15 <= 2.241657397)
Predict: 0.0
Else (feature 15 > 2.241657397)
If (feature 4 <= 0.5)
Predict: 1.0
Else (feature 4 > 0.5)
Predict: 0.0
Else (feature 32 > 51.5)
If (feature 20 <= 0.4474304455)
If (feature 15 <= 2.4902903465)
Predict: 1.0
Else (feature 15 > 2.4902903465)
Predict: 0.0
Else (feature 20 > 0.4474304455)
If (feature 30 <= 0.5)
Predict: 1.0
Else (feature 30 > 0.5)
Predict: 0.0
Else (feature 15 > 2.5904545455)
If (feature 35 <= 0.11862860350000001)
If (feature 20 <= 0.604100313)
If (feature 26 <= 0.0703273525)
Predict: 1.0
Else (feature 26 > 0.0703273525)
Predict: 0.0
Else (feature 20 > 0.604100313)
Predict: 0.0
Else (feature 35 > 0.11862860350000001)
If (feature 26 <= 0.0023895)
If (feature 27 <= 0.5)
Predict: 1.0
Else (feature 27 > 0.5)
Predict: 0.0
Else (feature 26 > 0.0023895)
Predict: 0.0
Else (feature 29 > 25.5)
If (feature 14 <= 0.5597875000000001)
If (feature 17 <= 0.1432107495)
If (feature 32 <= 84.5)
Predict: 0.0
Else (feature 32 > 84.5)
If (feature 32 <= 99.5)
Predict: 1.0
Else (feature 32 > 99.5)
Predict: 0.0
Else (feature 17 > 0.1432107495)
If (feature 15 <= 2.54534632)
If (feature 35 <= 0.1135839195)
If (feature 18 <= 0.071340821)
Predict: 0.0
Else (feature 18 > 0.071340821)
Predict: 1.0
Else (feature 35 > 0.1135839195)
Predict: 0.0
Else (feature 15 > 2.54534632)
If (feature 16 <= 0.842136969)
Predict: 0.0
Else (feature 16 > 0.842136969)
If (feature 35 <= 0.0704363165)
Predict: 0.0
Else (feature 35 > 0.0704363165)
Predict: 1.0
Else (feature 14 > 0.5597875000000001)
If (feature 9 <= 0.5)
If (feature 2 <= 0.5)
Predict: 0.0
Else (feature 2 > 0.5)
If (feature 17 <= 0.2244121985)
If (feature 31 <= 2302.5)
Predict: 1.0
Else (feature 31 > 2302.5)
Predict: 0.0
Else (feature 17 > 0.2244121985)
If (feature 32 <= 368.5)
Predict: 1.0
Else (feature 32 > 368.5)
Predict: 0.0
Else (feature 9 > 0.5)
If (feature 15 <= 2.9814851825)
If (feature 29 <= 54.5)
Predict: 0.0
Else (feature 29 > 54.5)
Predict: 1.0
Else (feature 15 > 2.9814851825)
Predict: 0.0
Else (feature 1 > 0.5)
If (feature 35 <= 0.10557280899999999)
If (feature 15 <= 2.6319293674999997)
If (feature 26 <= 0.0102600085)
If (feature 15 <= 1.982317179)
If (feature 20 <= 0.4708425435)
If (feature 14 <= 0.28213)
Predict: 0.0
Else (feature 14 > 0.28213)
Predict: 1.0
Else (feature 20 > 0.4708425435)
Predict: 0.0
Else (feature 15 > 1.982317179)
If (feature 21 <= 1.065E-4)
Predict: 0.0
Else (feature 21 > 1.065E-4)
Predict: 1.0
Else (feature 26 > 0.0102600085)
If (feature 25 <= 0.33658205350000003)
If (feature 35 <= 0.0833107985)
Predict: 1.0
Else (feature 35 > 0.0833107985)
If (feature 31 <= 1624.5)
Predict: 0.0
Else (feature 31 > 1624.5)
Predict: 1.0
Else (feature 25 > 0.33658205350000003)
Predict: 0.0
Else (feature 15 > 2.6319293674999997)
If (feature 23 <= 0.0459418095)
If (feature 29 <= 18.5)
Predict: 1.0
Else (feature 29 > 18.5)
If (feature 14 <= 0.0799874)
Predict: 0.0
Else (feature 14 > 0.0799874)
Predict: 1.0
Else (feature 23 > 0.0459418095)
If (feature 17 <= 0.4005988025)
If (feature 35 <= 0.09598409050000001)
If (feature 35 <= 0.090486137)
Predict: 0.0
Else (feature 35 > 0.090486137)
Predict: 1.0
Else (feature 35 > 0.09598409050000001)
Predict: 0.0
Else (feature 17 > 0.4005988025)
Predict: 1.0
Else (feature 35 > 0.10557280899999999)
If (feature 20 <= 0.43892298949999997)
If (feature 32 <= 384.5)
If (feature 25 <= 0.248139513)
If (feature 34 <= 0.1917453335)
Predict: 0.0
Else (feature 34 > 0.1917453335)
If (feature 15 <= 2.2038107279999997)
Predict: 1.0
Else (feature 15 > 2.2038107279999997)
Predict: 0.0
Else (feature 25 > 0.248139513)
If (feature 14 <= 0.2909735)
Predict: 1.0
Else (feature 14 > 0.2909735)
If (feature 23 <= 0.051961552)
Predict: 0.0
Else (feature 23 > 0.051961552)
Predict: 1.0
Else (feature 32 > 384.5)
If (feature 18 <= 0.093286713)
Predict: 1.0
Else (feature 18 > 0.093286713)
Predict: 0.0
Else (feature 20 > 0.43892298949999997)
If (feature 26 <= 7.9E-5)
Predict: 0.0
Else (feature 26 > 7.9E-5)
If (feature 20 <= 0.6911990685)
If (feature 23 <= 0.076318364)
If (feature 29 <= 36.5)
Predict: 1.0
Else (feature 29 > 36.5)
Predict: 0.0
Else (feature 23 > 0.076318364)
If (feature 14 <= 0.8499995)
Predict: 1.0
Else (feature 14 > 0.8499995)
Predict: 0.0
Else (feature 20 > 0.6911990685)
Predict: 1.0
Else (feature 0 > 0.5)
If (feature 35 <= 0.123543858)
If (feature 26 <= 0.0398114305)
If (feature 14 <= 0.8499995)
If (feature 14 <= 0.765962)
If (feature 32 <= 184.5)
If (feature 23 <= 0.0221340885)
Predict: 1.0
Else (feature 23 > 0.0221340885)
If (feature 17 <= 0.07201618500000001)
Predict: 0.0
Else (feature 17 > 0.07201618500000001)
Predict: 1.0
Else (feature 32 > 184.5)
If (feature 19 <= 0.0482599545)
If (feature 34 <= 0.092270351)
Predict: 0.0
Else (feature 34 > 0.092270351)
Predict: 1.0
Else (feature 19 > 0.0482599545)
Predict: 0.0
Else (feature 14 > 0.765962)
If (feature 26 <= 0.017958696500000003)
If (feature 14 <= 0.8432904999999999)
Predict: 0.0
Else (feature 14 > 0.8432904999999999)
If (feature 15 <= 1.971547115)
Predict: 1.0
Else (feature 15 > 1.971547115)
Predict: 0.0
Else (feature 26 > 0.017958696500000003)
If (feature 18 <= 4.04E-4)
Predict: 0.0
Else (feature 18 > 4.04E-4)
Predict: 1.0
Else (feature 14 > 0.8499995)
If (feature 25 <= 0.18025862)
Predict: 0.0
Else (feature 25 > 0.18025862)
Predict: 1.0
Else (feature 26 > 0.0398114305)
If (feature 29 <= 33.5)
If (feature 25 <= 0.1375444065)
If (feature 32 <= 214.5)
Predict: 1.0
Else (feature 32 > 214.5)
Predict: 0.0
Else (feature 25 > 0.1375444065)
If (feature 23 <= 0.1289247745)
If (feature 31 <= 1497.0)
Predict: 0.0
Else (feature 31 > 1497.0)
If (feature 35 <= 0.0263854075)
Predict: 0.0
Else (feature 35 > 0.0263854075)
Predict: 1.0
Else (feature 23 > 0.1289247745)
Predict: 0.0
Else (feature 29 > 33.5)
If (feature 19 <= 0.046103239000000004)
If (feature 17 <= 0.1148483135)
Predict: 0.0
Else (feature 17 > 0.1148483135)
Predict: 1.0
Else (feature 19 > 0.046103239000000004)
If (feature 23 <= 0.036442443)
If (feature 31 <= 10718.5)
If (feature 16 <= 0.499097473)
Predict: 0.0
Else (feature 16 > 0.499097473)
Predict: 1.0
Else (feature 31 > 10718.5)
Predict: 0.0
Else (feature 23 > 0.036442443)
If (feature 25 <= 0.24691742700000002)
If (feature 16 <= 0.6619797155)
Predict: 0.0
Else (feature 16 > 0.6619797155)
Predict: 1.0
Else (feature 25 > 0.24691742700000002)
Predict: 1.0
Else (feature 35 > 0.123543858)
If (feature 20 <= 0.494038735)
If (feature 19 <= 0.150085324)
Predict: 0.0
Else (feature 19 > 0.150085324)
Predict: 1.0
Else (feature 20 > 0.494038735)
If (feature 29 <= 33.5)
Predict: 1.0
Else (feature 29 > 33.5)
If (feature 23 <= 0.0285232765)
Predict: 0.0
Else (feature 23 > 0.0285232765)
If (feature 15 <= 2.078819647)
Predict: 1.0
Else (feature 15 > 2.078819647)
Predict: 0.0
Else (feature 3 > 0.5)
If (feature 29 <= 30.5)
If (feature 26 <= 0.0137855585)
If (feature 31 <= 3875.5)
Predict: 1.0
Else (feature 31 > 3875.5)
If (feature 29 <= 5.5)
Predict: 1.0
Else (feature 29 > 5.5)
If (feature 18 <= 0.08599717800000001)
Predict: 0.0
Else (feature 18 > 0.08599717800000001)
If (feature 18 <= 0.13203773600000002)
Predict: 1.0
Else (feature 18 > 0.13203773600000002)
Predict: 0.0
Else (feature 26 > 0.0137855585)
If (feature 14 <= 0.4769935)
If (feature 15 <= 1.5387538305000001)
Predict: 1.0
Else (feature 15 > 1.5387538305000001)
If (feature 19 <= 0.032281749)
Predict: 0.0
Else (feature 19 > 0.032281749)
If (feature 35 <= 0.099369468)
If (feature 27 <= 0.5)
If (feature 14 <= 0.3311615)
Predict: 1.0
Else (feature 14 > 0.3311615)
Predict: 0.0
Else (feature 27 > 0.5)
Predict: 1.0
Else (feature 35 > 0.099369468)
Predict: 0.0
Else (feature 14 > 0.4769935)
If (feature 18 <= 0.0119284765)
Predict: 0.0
Else (feature 18 > 0.0119284765)
If (feature 19 <= 0.0092025405)
If (feature 16 <= 0.350257732)
Predict: 1.0
Else (feature 16 > 0.350257732)
If (feature 29 <= 20.5)
Predict: 0.0
Else (feature 29 > 20.5)
Predict: 1.0
Else (feature 19 > 0.0092025405)
If (feature 29 <= 15.5)
If (feature 15 <= 2.6508928569999997)
If (feature 20 <= 0.338772014)
Predict: 1.0
Else (feature 20 > 0.338772014)
Predict: 0.0
Else (feature 15 > 2.6508928569999997)
If (feature 19 <= 0.0482599545)
Predict: 1.0
Else (feature 19 > 0.0482599545)
Predict: 0.0
Else (feature 29 > 15.5)
Predict: 0.0
Else (feature 29 > 30.5)
If (feature 16 <= 0.572688423)
If (feature 16 <= 0.554178497)
If (feature 26 <= 0.048132494)
If (feature 14 <= 0.7296955)
Predict: 0.0
Else (feature 14 > 0.7296955)
Predict: 1.0
Else (feature 26 > 0.048132494)
Predict: 0.0
Else (feature 16 > 0.554178497)
If (feature 14 <= 0.3016565)
Predict: 0.0
Else (feature 14 > 0.3016565)
Predict: 1.0
Else (feature 16 > 0.572688423)
Predict: 0.0
标签：1.0,Python,MLlib,feature,hadoop,Predict,0.0,15,Else 来源： https://www.cnblogs.com/zhuozige/p/12623301.html