Mise à l'échelle des fonctionnalités

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
 
df = pd.DataFrame({'WEIGHT': [15, 18, 12,10],
                   'PRICE': [1,3,2,5]},
                   index = ['Orange','Apple','Banana','Grape'])
print(df)
 
#Min-Max scaler
from sklearn.preprocessing import MinMaxScaler
scaler = MinMaxScaler()
df1 = pd.DataFrame(scaler.fit_transform(df),
                   columns=['WEIGHT','PRICE'],
                   index = ['Orange','Apple','Banana','Grape'])
ax = df.plot.scatter(x='WEIGHT', y='PRICE',color=['red','green','blue','yellow'],
                     marker = '*',s=80, label='BEFORE SCALING');
df1.plot.scatter(x='WEIGHT', y='PRICE', color=['red','green','blue','yellow'],
                 marker = 'o',s=60,label='AFTER SCALING', ax = ax);
plt.axhline(0, color='red',alpha=0.2)
plt.axvline(0, color='red',alpha=0.2);
 
#Standard Scaler
from sklearn.preprocessing import StandardScaler
scaler = StandardScaler()
df2 = pd.DataFrame(scaler.fit_transform(df),
                   columns=['WEIGHT','PRICE'],
                   index = ['Orange','Apple','Banana','Grape'])
ax = df.plot.scatter(x='WEIGHT', y='PRICE',color=['red','green','blue','yellow'],
                     marker = '*',s=80, label='BEFORE SCALING');
df2.plot.scatter(x='WEIGHT', y='PRICE', color=['red','green','blue','yellow'],
                 marker = 'o',s=60,label='AFTER SCALING', ax = ax)
plt.axhline(0, color='red',alpha=0.2)
plt.axvline(0, color='red',alpha=0.2);
 
#Max Abs Scaler
from sklearn.preprocessing import MaxAbsScaler
scaler = MaxAbsScaler()
df4 = pd.DataFrame(scaler.fit_transform(df),
                   columns=['WEIGHT','PRICE'],
                   index = ['Orange','Apple','Banana','Grape'])
ax = df.plot.scatter(x='WEIGHT', y='PRICE',color=['red','green','blue','yellow'],
                     marker = '*',s=80, label='BEFORE SCALING');
df4.plot.scatter(x='WEIGHT', y='PRICE', color=['red','green','blue','yellow'],
                 marker = 'o',s=60,label='AFTER SCALING', ax = ax)
plt.axhline(0, color='red',alpha=0.2)
plt.axvline(0, color='red',alpha=0.2);
 
#Robust Scaler
from sklearn.preprocessing import RobustScaler
scaler = RobustScaler()
df3 = pd.DataFrame(scaler.fit_transform(df),
                   columns=['WEIGHT','PRICE'],
                   index = ['Orange','Apple','Banana','Grape'])
ax = df.plot.scatter(x='WEIGHT', y='PRICE',color=['red','green','blue','yellow'],
                     marker = '*',s=80, label='BEFORE SCALING');
df3.plot.scatter(x='WEIGHT', y='PRICE', color=['red','green','blue','yellow'],
                 marker = 'o',s=60,label='AFTER SCALING', ax = ax)
plt.axhline(0, color='red',alpha=0.2)
plt.axvline(0, color='red',alpha=0.2);
 
dfr = pd.DataFrame({'WEIGHT': [15, 18, 12,10,50],
                   'PRICE': [1,3,2,5,20]},
                   index = ['Orange','Apple','Banana','Grape','Jackfruit'])
print(dfr)
 
from sklearn.preprocessing import StandardScaler
scaler = StandardScaler()
df21 = pd.DataFrame(scaler.fit_transform(dfr),
                   columns=['WEIGHT','PRICE'],
                   index = ['Orange','Apple','Banana','Grape','Jackfruit'])
ax = dfr.plot.scatter(x='WEIGHT', y='PRICE',color=['red','green','blue','yellow','black'],
                     marker = '*',s=80, label='BEFORE SCALING');
df21.plot.scatter(x='WEIGHT', y='PRICE', color=['red','green','blue','yellow','black'],
                 marker = 'o',s=60,label='STANDARD', ax = ax,figsize=(12,6))
from sklearn.preprocessing import RobustScaler
scaler = RobustScaler()
df31 = pd.DataFrame(scaler.fit_transform(dfr),
                   columns=['WEIGHT','PRICE'],
                   index = ['Orange','Apple','Banana','Grape','Jackfruit'])
df31.plot.scatter(x='WEIGHT', y='PRICE', color=['red','green','blue','yellow','black'],
                 marker = 'v',s=60,label='ROBUST', ax = ax,figsize=(12,6))
plt.axhline(0, color='red',alpha=0.2)
plt.axvline(0, color='red',alpha=0.2);
 
#Quantile Transformer Scaler
from sklearn.preprocessing import QuantileTransformer
scaler = QuantileTransformer()
df6 = pd.DataFrame(scaler.fit_transform(df),
                   columns=['WEIGHT','PRICE'],
                   index = ['Orange','Apple','Banana','Grape'])
ax = df.plot.scatter(x='WEIGHT', y='PRICE',color=['red','green','blue','yellow'],
                     marker = '*',s=80, label='BEFORE SCALING');
df6.plot.scatter(x='WEIGHT', y='PRICE', color=['red','green','blue','yellow'],
                 marker = 'o',s=60,label='AFTER SCALING', ax = ax,figsize=(6,4))
plt.axhline(0, color='red',alpha=0.2)
plt.axvline(0, color='red',alpha=0.2);
 
#Power Transformer Scaler
from sklearn.preprocessing import PowerTransformer
scaler = PowerTransformer(method='yeo-johnson')
df5 = pd.DataFrame(scaler.fit_transform(df),
                   columns=['WEIGHT','PRICE'],
                   index = ['Orange','Apple','Banana','Grape'])
ax = df.plot.scatter(x='WEIGHT', y='PRICE',color=['red','green','blue','yellow'],
                     marker = '*',s=80, label='BEFORE SCALING');
df5.plot.scatter(x='WEIGHT', y='PRICE', color=['red','green','blue','yellow'],
                 marker = 'o',s=60,label='AFTER SCALING', ax = ax)
plt.axhline(0, color='red',alpha=0.2)
plt.axvline(0, color='red',alpha=0.2);
 
#Unit Vector Scaler
df8=df.apply(lambda x : x/np.linalg.norm(x,1))
print(df8)
 
df9=df.apply(lambda x : x/np.linalg.norm(x,2))
print(df9)