Linear Regression

Task 1
In [1]:
import pandas as pd
import numpy as np
from matplotlib import pyplot as plt

In [2]:

df = pd.read_csv("homeprices.csv")
df.head()
Out[2]:

area price

0 2600 550000

1 3000 565000

2 3200 610000

3 3600 680000

4 4000 725000

In [3]:
inputs = df.drop("price",axis='columns')
inputs
Out[3]:

area

0 2600

1 3000

2 3200

3 3600

4 4000

In [4]:

target = df['price']

In [5]:
from sklearn.linear_model import LinearRegression
model = LinearRegression()

In [6]:
model.fit(inputs.values, target.values)

Out[6]:

▾ LinearRegression
LinearRegression()
In [7]:
model.predict([[3300]])
Out[7]:
array([628715.75342466])

In [8]:
model.predict([[5000]])
Out[8]:
array([859554.79452055])

Task 2
In [9]:
df1 = pd.read_csv("areas.csv")
df1.head()
Out[9]:

area

0 1000

1 1500

2 2300

3 3540

4 4120

In [10]:
p = model.predict(df1.values)
p
Out[10]:
array([ 316404.10958904, 384297.94520548, 492928.08219178,
661304.79452055, 740061.64383562, 799808.21917808,
926090.75342466, 650441.78082192, 825607.87671233,
492928.08219178, 1402705.47945205, 1348390.4109589 ,
1144708.90410959])

In [11]:
df1['prices'] = p
df.head()
Out[11]:

area price

0 2600 550000

1 3000 565000

2 3200 610000

3 3600 680000

4 4000 725000

In [12]:
df1.to_csv("Prices.csv",index=False)
In [13]:
df2 = pd.read_csv("Prices.csv")
df2.head()

Out[13]:

area prices

0 1000 316404.109589

1 1500 384297.945205

2 2300 492928.082192

3 3540 661304.794521

4 4120 740061.643836

Task 3
In [14]:
df3 = pd.read_csv("canada_pci.csv")
df3.head()
Out[14]:

year per capita income (US$)

0 1970 3399.299037

1 1971 3768.297935

2 1972 4251.175484

3 1973 4804.463248

4 1974 5576.514583

In [15]:
inputs1 = df3.drop("per capita income (US$)", axis='columns')
inputs1.head()
Out[15]:

year

0 1970

1 1971

2 1972

3 1973

4 1974

In [16]:
target1 = df3['per capita income (US$)']
target1.head()
Out[16]:
0 3399.299037
1 3768.297935
2 4251.175484
3 4804.463248
4 5576.514583
Name: per capita income (US$), dtype: float64

In [17]:
In [17]:
from sklearn.linear_model import LinearRegression
model1 = LinearRegression()

In [18]:
model1.fit(inputs1.values,target1.values)
Out[18]:

▾ LinearRegression
LinearRegression()

In [19]:
model1.predict([[2020]])
Out[19]:
array([41288.69409442])

In [20]:
model1.score(inputs1.values, target1.values)
Out[20]:
0.890916917957032

Day 2 Multiple Regression

In [22]:
df = pd.read_csv("homerates.csv")
df.head()
Out[22]:

area bedrooms age price

0 2600 3.0 20 550000

1 3000 4.0 15 565000

2 3200 NaN 18 610000

3 3600 3.0 30 595000

4 4000 5.0 8 760000

In [23]:
mn = df.bedrooms.mean()
mn
Out[23]:
4.2

In [24]:
df.bedrooms = df.bedrooms.fillna(mn)
df.head()
Out[24]:
 area bedrooms age price

0 2600 3.0 20 550000

1 3000 4.0 15 565000

2 3200 4.2 18 610000

3 3600 3.0 30 595000

4 4000 5.0 8 760000

In [25]:
inputs = df.drop('price',axis='columns')
inputs

Out[25]:

area bedrooms age

0 2600 3.0 20

1 3000 4.0 15

2 3200 4.2 18

3 3600 3.0 30

4 4000 5.0 8

5 4100 6.0 8

In [26]:
target = df['price']
target
Out[26]:
0 550000
1 565000
2 610000
3 595000
4 760000
5 810000
Name: price, dtype: int64

In [27]:
sns.pairplot(df)
Out[27]:
<seaborn.axisgrid.PairGrid at 0x22c1e3b25f0>
In [28]:
x = df.corr(numeric_only=True)
x
Out[28]:

area bedrooms age price

area 1.000000 0.740910 -0.445300 0.901476

bedrooms 0.740910 1.000000 -0.873162 0.910005

age -0.445300 -0.873162 1.000000 -0.734167

price 0.901476 0.910005 -0.734167 1.000000

In [29]:
sns.heatmap(x, annot=True, cmap='coolwarm')
Out[29]:
<Axes: >
In [30]:
from sklearn.linear_model import LinearRegression
model = LinearRegression()

In [31]:
model.fit(inputs.values,target.values)
Out[31]:

▾ LinearRegression
LinearRegression()

In [32]:
model.score(inputs.values, target.values)

Out[32]:
0.9540926625396438

In [33]:
model.predict([[300,3,40]])
Out[33]:
array([175825.67757211])

In [34]:
model.predict([[2500,4,5]])
Out[34]:
array([579906.16685223])

Task Multiple Regression

In [36]:
df1 = pd.read_csv("hiring.csv")
df1.head()
Out[36]:

test_score(out of interview_score(out of
experience salary($)
10) 10)

0 NaN 8.0 9 50000

1 NaN 8.0 6 45000

2 five 6.0 7 60000

3 two 10.0 10 65000

4 seven 9.0 6 70000

In [38]:
df1['test_score(out of 10)'] = df1['test_score(out of 10)'].fillna(df1['test_score(out of
10)'].mean())

In [39]:
df1.head()
Out[39]:

test_score(out of interview_score(out of
experience salary($)
10) 10)

0 zero 8.0 9 50000

1 zero 8.0 6 45000

2 five 6.0 7 60000

3 two 10.0 10 65000

4 seven 9.0 6 70000

In [40]:
df1.experience = df1.experience.apply(w2n.word_to_num)

In [41]:
df1.head()
Out[41]:

test_score(out of interview_score(out of
experience salary($)
10) 10)

0 0 8.0 9 50000

1 0 8.0 6 45000

2 5 6.0 7 60000

3 2 10.0 10 65000

4 7 9.0 6 70000

In [42]:
inputs1 = df1.drop("salary($)", axis = 'columns')
inputs1
Out[42]:

test_score(out of interview_score(out of
experience
10) 10)

0 0 8.000000 9

1 0 8.000000 6

2 5 6.000000 7

3 2 10.000000 10

4 7 9.000000 6

5 3 7.000000 10

6 10 7.857143 7

7 11 7.000000 8
In [43]:
target1 = df1['salary($)']
target1
Out[43]:
0 50000
1 45000
2 60000
3 65000
4 70000
5 62000
6 72000
7 80000
Name: salary($), dtype: int64

In [44]:
from sklearn.linear_model import LinearRegression
model1 = LinearRegression()

In [45]:
model1.fit(inputs1.values, target1.values)
Out[45]:

▾ LinearRegression
LinearRegression()

In [46]:
model1.predict([[2,9,6]])
Out[46]:
array([53290.89255945])

In [47]:
model1.predict([[12,10,10]])
Out[47]:
array([92268.07227784])

In [ ]: