Beginning Anomaly Detection Using

IRU

FRO

LQ

GIFROXPQV

LI

GI>FRO@GW\SH  REMHFW

HQFRGHG /DEHO(QFRGHU

HQFRGHGILW GI>FRO@

GI>FRO@ HQFRGHGWUDQVIRUP GI>FRO@

Figure 2-18.  Applying the label encoder to the columns with data values that are 

strings

Figure 2-19.  Looking at the first five entries of df after applying the label encoder

Chapter 2   traditional Methods of anoMaly deteCtion

---

45

With

df = df.iloc[np.random.permutation(len(df))]

you are randomly shuffling all the entries in the data set to avoid the problem of 

abnormal entries pooling in any one region of the data set.

With

df2 = df[:500000]

you are assigning the first 500,000 entries of df to a variable df2.

In the next line of code, labels = df2["label"], you assign the label column to  

the variable labels. Next, you assign the rest of the data frame to a variable named  

df_validate to create the validation data set with df_validate = df[500000:].

To split your data into the