Feature Engineering (2) - Data Preprocessing Interval Scaling Method

https://www.deeplearn.me/1383.html

 

Principle of interval scaling method

The most common is to use the maximum and minimum values ​​for processing. The processing formula is as follows

$$\mathrm{y=x-minmax-miny=x-minmax-min}$$

In the above formula, min represents the minimum value of the data, and max represents the maximum value of the data.

1. from sklearn.preprocessing importMinMaxScaler
2. tmp=MinMaxScaler().fit_transform(irisdata.data)
3. print tmp[0:5]

Partial results are as follows:
[[.22222222 0.625 0.06779661 0.04166667]
[0.16666667 0.06779661 0.04166667 .41666667]
[.11111111 0.5 0.05084746 0.04166667]
[.08333333 .45833333 .08474576 0.04166667]
[.19444444 .66666667 0.06779661 0.04166667]] Spark
scaling method in interval

1. >>>from pyspark.mllib.linalg importVectors
2. >>>from pyspark.sql importSQLContext
3. >>>sqlContext=SQLContext(sc)
4. >>>df = sqlContext.createDataFrame([(Vectors.dense([0.0]),),(Vectors.dense([2.0]),)],["a"])
5. >>> mmScaler =MinMaxScaler(inputCol="a", outputCol="scaled")
6. >>> model = mmScaler.fit(df)
7. >>> model.transform(df).show()
8. +-----+------+
9. | a|scaled|
10. +-----+------+
11. |[0.0]| [0.0]|
12. |[2.0]| [1.0]|