PySpark 초보자를 위한 튜토리얼: 예시를 통해 학습

sudo yum update -y
sudo yum install -y docker
sudo service docker start
sudo user-mod -a -G docker ec2-user
exit

## Spark
docker run -v ~/work:/home/jovyan/work -d -p 8888:8888 jupyter/pyspark-notebook

## Allow preserving Jupyter notebook
sudo chown 1000 ~/work

## Install tree to see our working directory next
sudo yum install -y tree

docker ps

cd anaconda3
touch hello-spark.yml
vi hello-spark.yml

cd C:\Users\Admin\Anaconda3
echo.>hello-spark.yml
notepad hello-spark.yml

name: hello-spark 
    dependencies:
    
    - python=3.6
    - jupyter
    - ipython
    - numpy
    - numpy-base
    - pandas
    - py4j
    - pyspark
    - pytz

conda env create -f hello-spark.yml

conda env list

Activate hello-spark

source activate hello-spark

activate hello-spark

import pyspark
from pyspark import SparkContext
sc =SparkContext()

brew tap caskroom/versions
brew cask install java8

 conda env remove -n hello-spark -y

import pyspark
from pyspark import SparkContext
sc =SparkContext()

nums= sc.parallelize([1,2,3,4])

nums.take(1)

[1]

squared = nums.map(lambda x: x*x).collect()
for num in squared:
    print('%i ' % (num))

1 
4 
9 
16

from pyspark.sql import Row
from pyspark.sql import SQLContext

sqlContext = SQLContext(sc)

[('John',19),('Smith',29),('Adam',35),('Henry',50)]

rdd = sc.parallelize(list_p)

rdd.map(lambda x: Row(name=x[0], age=int(x[1])))

sqlContext.createDataFrame(ppl)
list_p = [('John',19),('Smith',29),('Adam',35),('Henry',50)]
rdd = sc.parallelize(list_p)
ppl = rdd.map(lambda x: Row(name=x[0], age=int(x[1])))
DF_ppl = sqlContext.createDataFrame(ppl)

DF_ppl.printSchema()
root
 |-- age: long (nullable = true)
 |-- name: string (nullable = true)

#from pyspark.sql import SQLContext
url = "https://raw.githubusercontent.com/guru99-edu/R-Programming/master/adult_data.csv"
from pyspark import SparkFiles
sc.addFile(url)
sqlContext = SQLContext(sc)

df = sqlContext.read.csv(SparkFiles.get("adult_data.csv"), header=True, inferSchema= True)

df.printSchema()
root
 |-- age: integer (nullable = true)
 |-- workclass: string (nullable = true)
 |-- fnlwgt: integer (nullable = true)
 |-- education: string (nullable = true)
 |-- education_num: integer (nullable = true)
 |-- marital: string (nullable = true)
 |-- occupation: string (nullable = true)
 |-- relationship: string (nullable = true)
 |-- race: string (nullable = true)
 |-- sex: string (nullable = true)
 |-- capital_gain: integer (nullable = true)
 |-- capital_loss: integer (nullable = true)
 |-- hours_week: integer (nullable = true)
 |-- native_country: string (nullable = true)
 |-- label: string (nullable = true)

df.show(5, truncate = False)

+---+----------------+------+---------+-------------+------------------+-----------------+-------------+-----+------+------------+------------+----------+--------------+-----+
|age|workclass       |fnlwgt|education|education_num|marital           |occupation       |relationship |race |sex   |capital_gain|capital_loss|hours_week|native_country|label|
+---+----------------+------+---------+-------------+------------------+-----------------+-------------+-----+------+------------+------------+----------+--------------+-----+
|39 |State-gov       |77516 |Bachelors|13           |Never-married     |Adm-clerical     |Not-in-family|White|Male  |2174        |0           |40        |United-States |<=50K|
|50 |Self-emp-not-inc|83311 |Bachelors|13           |Married-civ-spouse|Exec-managerial  |Husband      |White|Male  |0           |0           |13        |United-States |<=50K|
|38 |Private         |215646|HS-grad  |9            |Divorced          |Handlers-cleaners|Not-in-family|White|Male  |0           |0           |40        |United-States |<=50K|
|53 |Private         |234721|11th     |7            |Married-civ-spouse|Handlers-cleaners|Husband      |Black|Male  |0           |0           |40        |United-States |<=50K|
|28 |Private         |338409|Bachelors|13           |Married-civ-spouse|Prof-specialty   |Wife         |Black|Female|0           |0           |40        |Cuba          |<=50K|
+---+----------------+------+---------+-------------+------------------+-----------------+-------------+-----+------+------------+------------+----------+--------------+-----+
only showing top 5 rows

df_string = sqlContext.read.csv(SparkFiles.get("adult.csv"), header=True, inferSchema=  False)
df_string.printSchema()
root
 |-- age: string (nullable = true)
 |-- workclass: string (nullable = true)
 |-- fnlwgt: string (nullable = true)
 |-- education: string (nullable = true)
 |-- education_num: string (nullable = true)
 |-- marital: string (nullable = true)
 |-- occupation: string (nullable = true)
 |-- relationship: string (nullable = true)
 |-- race: string (nullable = true)
 |-- sex: string (nullable = true)
 |-- capital_gain: string (nullable = true)
 |-- capital_loss: string (nullable = true)
 |-- hours_week: string (nullable = true)
 |-- native_country: string (nullable = true)
 |-- label: string (nullable = true)

# Import all from `sql.types`
from pyspark.sql.types import *

# Write a custom function to convert the data type of DataFrame columns
def convertColumn(df, names, newType):
    for name in names: 
        df = df.withColumn(name, df[name].cast(newType))
    return df 
# List of continuous features
CONTI_FEATURES  = ['age', 'fnlwgt','capital_gain', 'education_num', 'capital_loss', 'hours_week']
# Convert the type
df_string = convertColumn(df_string, CONTI_FEATURES, FloatType())
# Check the dataset
df_string.printSchema()
root
 |-- age: float (nullable = true)
 |-- workclass: string (nullable = true)
 |-- fnlwgt: float (nullable = true)
 |-- education: string (nullable = true)
 |-- education_num: float (nullable = true)
 |-- marital: string (nullable = true)
 |-- occupation: string (nullable = true)
 |-- relationship: string (nullable = true)
 |-- race: string (nullable = true)
 |-- sex: string (nullable = true)
 |-- capital_gain: float (nullable = true)
 |-- capital_loss: float (nullable = true)
 |-- hours_week: float (nullable = true)
 |-- native_country: string (nullable = true)
 |-- label: string (nullable = true)

from pyspark.ml.feature import StringIndexer
#stringIndexer = StringIndexer(inputCol="label", outputCol="newlabel")
#model = stringIndexer.fit(df)
#df = model.transform(df)
df.printSchema()

df.select('age','fnlwgt').show(5)

+---+------+
|age|fnlwgt|
+---+------+
| 39| 77516|
| 50| 83311|
| 38|215646|
| 53|234721|
| 28|338409|
+---+------+
only showing top 5 rows

df.groupBy("education").count().sort("count",ascending=True).show()

+------------+-----+
|   education|count|
+------------+-----+
|   Preschool|   51|
|     1st-4th|  168|
|     5th-6th|  333|
|   Doctorate|  413|
|        12th|  433|
|         9th|  514|
| Prof-school|  576|
|     7th-8th|  646|
|        10th|  933|
|  Assoc-acdm| 1067|
|        11th| 1175|
|   Assoc-voc| 1382|
|     Masters| 1723|
|   Bachelors| 5355|
|Some-college| 7291|
|     HS-grad|10501|
+------------+-----+

df.describe().show()

+-------+------------------+-----------+------------------+------------+-----------------+--------+----------------+------------+------------------+------+------------------+----------------+------------------+--------------+-----+
|summary|               age|  workclass|            fnlwgt|   education|    education_num| marital|      occupation|relationship|              race|   sex|      capital_gain|    capital_loss|        hours_week|native_country|label|
+-------+------------------+-----------+------------------+------------+-----------------+--------+----------------+------------+------------------+------+------------------+----------------+------------------+--------------+-----+
|  count|             32561|      32561|             32561|       32561|            32561|   32561|           32561|       32561|             32561| 32561|             32561|           32561|             32561|         32561|32561|
|   mean| 38.58164675532078|       null|189778.36651208502|        null| 10.0806793403151|    null|            null|        null|              null|  null|1077.6488437087312| 87.303829734959|40.437455852092995|          null| null|
| stddev|13.640432553581356|       null|105549.97769702227|        null|2.572720332067397|    null|            null|        null|              null|  null| 7385.292084840354|402.960218649002|12.347428681731838|          null| null|
|    min|                17|          ?|             12285|        10th|                1|Divorced|               ?|     Husband|Amer-Indian-Eskimo|Female|                 0|               0|                 1|             ?|<=50K|
|    max|                90|Without-pay|           1484705|Some-college|               16| Widowed|Transport-moving|        Wife|             White|  Male|             99999|            4356|                99|    Yugoslavia| >50K|
+-------+------------------+-----------+------------------+------------+-----------------+--------+----------------+------------+------------------+------+------------------+----------------+------------------+--------------+-----+

df.describe('capital_gain').show()

+-------+------------------+
|summary|      capital_gain|
+-------+------------------+
|  count|             32561|
|   mean|1077.6488437087312|
| stddev| 7385.292084840354|
|    min|                 0|
|    max|             99999|
+-------+------------------+

df.crosstab('age', 'label').sort("age_label").show()

+---------+-----+----+
|age_label|<=50K|>50K|
+---------+-----+----+
|       17|  395|   0|
|       18|  550|   0|
|       19|  710|   2|
|       20|  753|   0|
|       21|  717|   3|
|       22|  752|  13|
|       23|  865|  12|
|       24|  767|  31|
|       25|  788|  53|
|       26|  722|  63|
|       27|  754|  81|
|       28|  748| 119|
|       29|  679| 134|
|       30|  690| 171|
|       31|  705| 183|
|       32|  639| 189|
|       33|  684| 191|
|       34|  643| 243|
|       35|  659| 217|
|       36|  635| 263|
+---------+-----+----+
only showing top 20 rows

df.drop('education_num').columns

['age',
 'workclass',
 'fnlwgt',
 'education',
 'marital',
 'occupation',
 'relationship',
 'race',
 'sex',
 'capital_gain',
 'capital_loss',
 'hours_week',
 'native_country',
 'label']

df.filter(df.age > 40).count()

df.groupby('marital').agg({'capital_gain': 'mean'}).show()

+--------------------+------------------+
|             marital| avg(capital_gain)|
+--------------------+------------------+
|           Separated| 535.5687804878049|
|       Never-married|376.58831788823363|
|Married-spouse-ab...| 653.9832535885167|
|            Divorced| 728.4148098131893|
|             Widowed| 571.0715005035247|
|   Married-AF-spouse| 432.6521739130435|
|  Married-civ-spouse|1764.8595085470085|
+--------------------+------------------+

from pyspark.sql.functions import *

# 1 Select the column
age_square = df.select(col("age")**2)

# 2 Apply the transformation and add it to the DataFrame
df = df.withColumn("age_square", col("age")**2)

df.printSchema()
root
 |-- age: integer (nullable = true)
 |-- workclass: string (nullable = true)
 |-- fnlwgt: integer (nullable = true)
 |-- education: string (nullable = true)
 |-- education_num: integer (nullable = true)
 |-- marital: string (nullable = true)
 |-- occupation: string (nullable = true)
 |-- relationship: string (nullable = true)
 |-- race: string (nullable = true)
 |-- sex: string (nullable = true)
 |-- capital_gain: integer (nullable = true)
 |-- capital_loss: integer (nullable = true)
 |-- hours_week: integer (nullable = true)
 |-- native_country: string (nullable = true)
 |-- label: string (nullable = true)
 |-- age_square: double (nullable = true)

COLUMNS = ['age', 'age_square', 'workclass', 'fnlwgt', 'education', 'education_num', 'marital',
           'occupation', 'relationship', 'race', 'sex', 'capital_gain', 'capital_loss',
           'hours_week', 'native_country', 'label']
df = df.select(COLUMNS)
df.first()

Row(age=39, age_square=1521.0, workclass='State-gov', fnlwgt=77516, education='Bachelors', education_num=13, marital='Never-married', occupation='Adm-clerical', relationship='Not-in-family', race='White', sex='Male', capital_gain=2174, capital_loss=0, hours_week=40, native_country='United-States', label='<=50K')

df.filter(df.native_country == 'Holand-Netherlands').count()
df.groupby('native_country').agg({'native_country': 'count'}).sort(asc("count(native_country)")).show()

+--------------------+---------------------+
|      native_country|count(native_country)|
+--------------------+---------------------+
|  Holand-Netherlands|                    1|
|            Scotland|                   12|
|             Hungary|                   13|
|            Honduras|                   13|
|Outlying-US(Guam-...|                   14|
|          Yugoslavia|                   16|
|            Thailand|                   18|
|                Laos|                   18|
|            Cambodia|                   19|
|     Trinadad&Tobago|                   19|
|                Hong|                   20|
|             Ireland|                   24|
|             Ecuador|                   28|
|              Greece|                   29|
|              France|                   29|
|                Peru|                   31|
|           Nicaragua|                   34|
|            Portugal|                   37|
|                Iran|                   43|
|               Haiti|                   44|
+--------------------+---------------------+
only showing top 20 rows

df_remove = df.filter(df.native_country !=	'Holand-Netherlands')

StringIndexer(inputCol="workclass", outputCol="workclass_encoded")

model = stringIndexer.fit(df)		
`indexed = model.transform(df)``

OneHotEncoder(dropLast=False, inputCol="workclassencoded", outputCol="workclassvec")

### Example encoder
from pyspark.ml.feature import StringIndexer, OneHotEncoder, VectorAssembler

stringIndexer = StringIndexer(inputCol="workclass", outputCol="workclass_encoded")
model = stringIndexer.fit(df)
indexed = model.transform(df)
encoder = OneHotEncoder(dropLast=False, inputCol="workclass_encoded", outputCol="workclass_vec")
encoded = encoder.transform(indexed)
encoded.show(2)

+---+----------+----------------+------+---------+-------------+------------------+---------------+-------------+-----+----+------------+------------+----------+--------------+-----+-----------------+-------------+
|age|age_square|       workclass|fnlwgt|education|education_num|           marital|     occupation| relationship| race| sex|capital_gain|capital_loss|hours_week|native_country|label|workclass_encoded|workclass_vec|
+---+----------+----------------+------+---------+-------------+------------------+---------------+-------------+-----+----+------------+------------+----------+--------------+-----+-----------------+-------------+
| 39|    1521.0|       State-gov| 77516|Bachelors|           13|     Never-married|   Adm-clerical|Not-in-family|White|Male|        2174|           0|        40| United-States|<=50K|              4.0|(9,[4],[1.0])|
| 50|    2500.0|Self-emp-not-inc| 83311|Bachelors|           13|Married-civ-spouse|Exec-managerial|      Husband|White|Male|           0|           0|        13| United-States|<=50K|              1.0|(9,[1],[1.0])|
+---+----------+----------------+------+---------+-------------+------------------+---------------+-------------+-----+----+------------+------------+----------+--------------+-----+-----------------+-------------+
only showing top 2 rows

from pyspark.ml import Pipeline
from pyspark.ml.feature import OneHotEncoderEstimator
CATE_FEATURES = ['workclass', 'education', 'marital', 'occupation', 'relationship', 'race', 'sex', 'native_country']
stages = [] # stages in our Pipeline
for categoricalCol in CATE_FEATURES:
    stringIndexer = StringIndexer(inputCol=categoricalCol, outputCol=categoricalCol + "Index")
    encoder = OneHotEncoderEstimator(inputCols=[stringIndexer.getOutputCol()],
                                     outputCols=[categoricalCol + "classVec"])
    stages += [stringIndexer, encoder]

# Convert label into label indices using the StringIndexer
label_stringIdx =  StringIndexer(inputCol="label", outputCol="newlabel")
stages += [label_stringIdx]

assemblerInputs = [c + "classVec" for c in CATE_FEATURES] + CONTI_FEATURES

assembler = VectorAssembler(inputCols=assemblerInputs, outputCol="features")stages += [assembler]

# Create a Pipeline.
pipeline = Pipeline(stages=stages)
pipelineModel = pipeline.fit(df_remove)
model = pipelineModel.transform(df_remove)

model.take(1)

[Row(age=39, age_square=1521.0, workclass='State-gov', fnlwgt=77516, education='Bachelors', education_num=13, marital='Never-married', occupation='Adm-clerical', relationship='Not-in-family', race='White', sex='Male', capital_gain=2174, capital_loss=0, hours_week=40, native_country='United-States', label='<=50K', workclassIndex=4.0, workclassclassVec=SparseVector(8, {4: 1.0}), educationIndex=2.0, educationclassVec=SparseVector(15, {2: 1.0}), maritalIndex=1.0, maritalclassVec=SparseVector(6, {1: 1.0}), occupationIndex=3.0, occupationclassVec=SparseVector(14, {3: 1.0}), relationshipIndex=1.0, relationshipclassVec=SparseVector(5, {1: 1.0}), raceIndex=0.0, raceclassVec=SparseVector(4, {0: 1.0}), sexIndex=0.0, sexclassVec=SparseVector(1, {0: 1.0}), native_countryIndex=0.0, native_countryclassVec=SparseVector(40, {0: 1.0}), newlabel=0.0, features=SparseVector(99, {4: 1.0, 10: 1.0, 24: 1.0, 32: 1.0, 44: 1.0, 48: 1.0, 52: 1.0, 53: 1.0, 93: 39.0, 94: 77516.0, 95: 2174.0, 96: 13.0, 98: 40.0}))]

from pyspark.ml.linalg import DenseVector
input_data = model.rdd.map(lambda x: (x["newlabel"], DenseVector(x["features"])))

df_train = sqlContext.createDataFrame(input_data, ["label", "features"])

df_train.show(2)

+-----+--------------------+
|label|            features|
+-----+--------------------+
|  0.0|[0.0,0.0,0.0,0.0,...|
|  0.0|[0.0,1.0,0.0,0.0,...|
+-----+--------------------+
only showing top 2 rows

# Split the data into train and test sets
train_data, test_data = df_train.randomSplit([.8,.2],seed=1234)

train_data.groupby('label').agg({'label': 'count'}).show()

+-----+------------+
|label|count(label)|
+-----+------------+
|  0.0|       19698|
|  1.0|        6263|
+-----+------------+

test_data.groupby('label').agg({'label': 'count'}).show()

+-----+------------+
|label|count(label)|
+-----+------------+
|  0.0|        5021|
|  1.0|        1578|
+-----+------------+

# Import `LinearRegression`
from pyspark.ml.classification import LogisticRegression

# Initialize `lr`
lr = LogisticRegression(labelCol="label",
                        featuresCol="features",
                        maxIter=10,
                        regParam=0.3)

# Fit the data to the model
linearModel = lr.fit(train_data)

# Print the coefficients and intercept for logistic regression
print("Coefficients: " + str(linearModel.coefficients))
print("Intercept: " + str(linearModel.intercept))

Coefficients: [-0.0678914665262,-0.153425526813,-0.0706009536407,-0.164057586562,-0.120655298528,0.162922330862,0.149176870438,-0.626836362611,-0.193483661541,-0.0782269980838,0.222667203836,0.399571096381,-0.0222024341804,-0.311925857859,-0.0434497788688,-0.306007744328,-0.41318209688,0.547937504247,-0.395837350854,-0.23166535958,0.618743906733,-0.344088614546,-0.385266881369,0.317324463006,-0.350518889186,-0.201335923138,-0.232878560088,-0.13349278865,-0.119760542498,0.17500602491,-0.0480968101118,0.288484253943,-0.116314616745,0.0524163478063,-0.300952624551,-0.22046421474,-0.16557996579,-0.114676231939,-0.311966431453,-0.344226119233,0.105530129507,0.152243047814,-0.292774545497,0.263628334433,-0.199951374076,-0.30329422583,-0.231087515178,0.418918551,-0.0565930184279,-0.177818073048,-0.0733236680663,-0.267972912252,0.168491215697,-0.12181255723,-0.385648075442,-0.202101794517,0.0469791640782,-0.00842850210625,-0.00373211448629,-0.259296141281,-0.309896554133,-0.168434409756,-0.11048086026,0.0280647963877,-0.204187030092,-0.414392623536,-0.252806580669,0.143366465705,-0.516359222663,-0.435627370849,-0.301949286524,0.0878249035894,-0.210951740965,-0.621417928742,-0.099445190784,-0.232671473401,-0.1077745606,-0.360429419703,-0.420362959052,-0.379729467809,-0.395186242741,0.0826401853838,-0.280251589972,0.187313505214,-0.20295228799,-0.431177064626,0.149759018379,-0.107114299614,-0.319314858424,0.0028450133235,-0.651220387649,-0.327918792207,-0.143659581445,0.00691075160413,8.38517628783e-08,2.18856717378e-05,0.0266701216268,0.000231075966823,0.00893832698698]
Intercept: -1.9884177974805692

# Make predictions on test data using the transform() method.
predictions = linearModel.transform(test_data)

predictions.printSchema()
root
 |-- label: double (nullable = true)
 |-- features: vector (nullable = true)
 |-- rawPrediction: vector (nullable = true)
 |-- probability: vector (nullable = true)
 |-- prediction: double (nullable = false)

selected = predictions.select("label", "prediction", "probability")
selected.show(20)

+-----+----------+--------------------+
|label|prediction|         probability|
+-----+----------+--------------------+
|  0.0|       0.0|[0.91560704124179...|
|  0.0|       0.0|[0.92812140213994...|
|  0.0|       0.0|[0.92161406774159...|
|  0.0|       0.0|[0.96222760777142...|
|  0.0|       0.0|[0.66363283056957...|
|  0.0|       0.0|[0.65571324475477...|
|  0.0|       0.0|[0.73053376932829...|
|  0.0|       1.0|[0.31265053873570...|
|  0.0|       0.0|[0.80005907577390...|
|  0.0|       0.0|[0.76482251301640...|
|  0.0|       0.0|[0.84447301189069...|
|  0.0|       0.0|[0.75691912026619...|
|  0.0|       0.0|[0.60902504096722...|
|  0.0|       0.0|[0.80799228385509...|
|  0.0|       0.0|[0.87704364852567...|
|  0.0|       0.0|[0.83817652582377...|
|  0.0|       0.0|[0.79655423248500...|
|  0.0|       0.0|[0.82712311232246...|
|  0.0|       0.0|[0.81372823882016...|
|  0.0|       0.0|[0.59687710752201...|
+-----+----------+--------------------+
only showing top 20 rows

cm = predictions.select("label", "prediction")

cm.groupby('label').agg({'label': 'count'}).show()

+-----+------------+
|label|count(label)|
+-----+------------+
|  0.0|        5021|
|  1.0|        1578|
+-----+------------+

cm.groupby('prediction').agg({'prediction': 'count'}).show()

+----------+-----------------+
|prediction|count(prediction)|
+----------+-----------------+
|       0.0|             5982|
|       1.0|              617|
+----------+-----------------+

cm.filter(cm.label == cm.prediction).count() / cm.count()

def accuracy_m(model): 
    predictions = model.transform(test_data)
    cm = predictions.select("label", "prediction")
    acc = cm.filter(cm.label == cm.prediction).count() / cm.count()
    print("Model accuracy: %.3f%%" % (acc * 100)) 
accuracy_m(model = linearModel)
Model accuracy: 82.376%

### Use ROC 
from pyspark.ml.evaluation import BinaryClassificationEvaluator

# Evaluate model
evaluator = BinaryClassificationEvaluator(rawPredictionCol="rawPrediction")
print(evaluator.evaluate(predictions))
print(evaluator.getMetricName())

print(evaluator.evaluate(predictions))

from pyspark.ml.tuning import ParamGridBuilder, CrossValidator

# Create ParamGrid for Cross Validation
paramGrid = (ParamGridBuilder()
             .addGrid(lr.regParam, [0.01, 0.5])
             .build())

from time import *
start_time = time()

# Create 5-fold CrossValidator
cv = CrossValidator(estimator=lr,
                    estimatorParamMaps=paramGrid,
                    evaluator=evaluator, numFolds=5)

# Run cross validations
cvModel = cv.fit(train_data)
# likely take a fair amount of time
end_time = time()
elapsed_time = end_time - start_time
print("Time to train model: %.3f seconds" % elapsed_time)

accuracy_m(model = cvModel)
Model accuracy: 85.316%

bestModel = cvModel.bestModel
bestModel.extractParamMap()

{Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='aggregationDepth', doc='suggested depth for treeAggregate (>= 2)'): 2,
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='elasticNetParam', doc='the ElasticNet mixing parameter, in range [0, 1]. For alpha = 0, the penalty is an L2 penalty. For alpha = 1, it is an L1 penalty'): 0.0,
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='family', doc='The name of family which is a description of the label distribution to be used in the model. Supported options: auto, binomial, multinomial.'): 'auto',
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='featuresCol', doc='features column name'): 'features',
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='fitIntercept', doc='whether to fit an intercept term'): True,
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='labelCol', doc='label column name'): 'label',
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='maxIter', doc='maximum number of iterations (>= 0)'): 10,
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='predictionCol', doc='prediction column name'): 'prediction',
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='probabilityCol', doc='Column name for predicted class conditional probabilities. Note: Not all models output well-calibrated probability estimates! These probabilities should be treated as confidences, not precise probabilities'): 'probability',
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='rawPredictionCol', doc='raw prediction (a.k.a. confidence) column name'): 'rawPrediction',
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='regParam', doc='regularization parameter (>= 0)'): 0.01,
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='standardization', doc='whether to standardize the training features before fitting the model'): True,
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='threshold', doc='threshold in binary classification prediction, in range [0, 1]'): 0.5,
 Param(parent='LogisticRegression_4d8f8ce4d6a02d8c29a0', name='tol', doc='the convergence tolerance for iterative algorithms (>= 0)'): 1e-06}