CODEWITHSUNDEEP
pythonpandas
siddharth kshirsagar
siddharth kshirsagar

Reputation: 113

How to make a custom confusion matrix using pandas

By using .value_counts() on my dataframe I got a series as follows:

Sample data:

In_A_Not_In_B    116
In_C_Not_In_B    104
In_A_Not_In_C    100
In_C_Not_In_A    85

I need to create a matrix out of the above data as shown below. The index names have to be fetched from the column value.

Expected output:

      NOT IN
      A  B  C
   A NA 116 100
IN C 85 104 NA

So far I have tried with .pivot_table(), .MultiIndex() with no luck.

Upvotes: 4

Views: 439

Answers (1)

zipa
zipa

Reputation: 27879

Let's say that your result of value_counts is stored in variable df, then:

df = df.reset_index()
df['In'] = df['index'].str[3]
df['Not In'] = df['index'].str[-1]
df.pivot(index='In', columns='Not In', values=0)
# Not In     A      B      C
# In                        
# A        NaN  116.0  100.0
# C       85.0  104.0    NaN

Upvotes: 4

Related Questions