03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [1]:  import pandas as pd

In [2]:  from matplotlib import pyplot as plt

In [3]:  import seaborn as sns

In [4]:  poke =pd.read_csv(r"D:\smriti iitr and tarot material\data science and

(https://getlin

In [5]:  poke

Out[5]:
Type Sp. Sp.
# Name Type 1 Total HP Attack Defense Speed
2 Atk Def

0 1 Bulbasaur Grass Poison 318 45 49 49 65 65 45

1 2 Ivysaur Grass Poison 405 60 62 63 80 80 60

2 3 Venusaur Grass Poison 525 80 82 83 100 100 80

VenusaurMega
3 3 Grass Poison 625 80 100 123 122 120 80
Venusaur

4 4 Charmander Fire NaN 309 39 52 43 60 50 65

... ... ... ... ... ... ... ... ... ... ... ...

795 719 Diancie Rock Fairy 600 50 100 150 100 150 50

DiancieMega
796 719 Rock Fairy 700 50 160 110 160 110 110
Diancie

HoopaHoopa
797 720 Psychic Ghost 600 80 110 60 150 130 70
Confined

HoopaHoopa
798 720 Psychic Dark 680 80 160 60 170 130 80
Unbound

799 721 Volcanion Fire Water 600 80 110 120 130 90 70

800 rows × 13 columns

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 1/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [6]:  poke.head()

Out[6]:
Type Type Sp. Sp.
# Name Total HP Attack Defense Speed Gene
1 2 Atk Def

0 1 Bulbasaur Grass Poison 318 45 49 49 65 65 45

1 2 Ivysaur Grass Poison 405 60 62 63 80 80 60

2 3 Venusaur Grass Poison 525 80 82 83 100 100 80

VenusaurMega
3 3 Grass Poison 625 80 100 123 122 120 80
Venusaur
(https://getlin
4 4 Charmander Fire NaN 309 39 52 43 60 50 65

In [7]:  poke.rename(columns= {'Type 1':'Primary_type', 'Type 2': 'Secondary_typ

In [8]:  poke.head()

Out[8]:
Sp. S
# Name Primary_type Secondary_type Total HP Attack Defense
Atk D

0 1 Bulbasaur Grass Poison 318 45 49 49 65 6

1 2 Ivysaur Grass Poison 405 60 62 63 80 8

2 3 Venusaur Grass Poison 525 80 82 83 100 10

VenusaurMega
3 3 Grass Poison 625 80 100 123 122 12
Venusaur

4 4 Charmander Fire NaN 309 39 52 43 60 5

In [9]:  poke['Primary_type']== 'Grass'

Out[9]: 0 True
1 True
2 True
3 True
4 False
...
795 False
796 False
797 False
798 False
799 False
Name: Primary_type, Length: 800, dtype: bool

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 2/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [10]:  #to extract grass type pokemon, true values

​
poke[ poke['Primary_type']== 'Grass']

Out[10]:
S
# Name Primary_type Secondary_type Total HP Attack Defense
At

0 1 Bulbasaur Grass Poison 318 45 49 49 6

1 2 Ivysaur Grass Poison 405 60 62 63 8

2 3 Venusaur Grass Poison 525 80 82 83 10

(https://getlin
VenusaurMega
3 3 Grass Poison 625 80 100 123 12
Venusaur

48 43 Oddish Grass Poison 320 45 50 55 7

... ... ... ... ... ... ... ... ...

718 650 Chespin Grass NaN 313 56 61 65 4

719 651 Quilladin Grass NaN 405 61 78 95 5

720 652 Chesnaught Grass Fighting 530 88 107 122 7

740 672 Skiddo Grass NaN 350 66 65 48 6

741 673 Gogoat Grass NaN 531 123 100 62 9

70 rows × 13 columns

In [11]:  #store in new object called grass_pokemon

grass_pokemon = poke[ poke['Primary_type']== 'Grass']

In [12]:  grass_pokemon.head()

Out[12]:
Sp.
# Name Primary_type Secondary_type Total HP Attack Defense
Atk

0 1 Bulbasaur Grass Poison 318 45 49 49 65

1 2 Ivysaur Grass Poison 405 60 62 63 80

2 3 Venusaur Grass Poison 525 80 82 83 100

VenusaurMega
3 3 Grass Poison 625 80 100 123 122
Venusaur

48 43 Oddish Grass Poison 320 45 50 55 75

In [13]:  water_pokemon = poke[ poke['Primary_type']== 'Water']

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 3/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [14]:  water_pokemon.head()

Out[14]:
Sp.
# Name Primary_type Secondary_type Total HP Attack Defense
Atk

9 7 Squirtle Water NaN 314 44 48 65 50

10 8 Wartortle Water NaN 405 59 63 80 65

11 9 Blastoise Water NaN 530 79 83 100 85

BlastoiseMega
12 9 Water NaN 630 79 103 120 135
Blastoise
(https://getlin
59 54 Psyduck Water NaN 320 50 52 48 65

In [15]:  fire_pokemon = poke[ poke['Primary_type']== 'Fire']

In [16]:  fire_pokemon.head()

Out[16]:
Sp. S
# Name Primary_type Secondary_type Total HP Attack Defense
Atk D

4 4 Charmander Fire NaN 309 39 52 43 60

5 5 Charmeleon Fire NaN 405 58 64 58 80

6 6 Charizard Fire Flying 534 78 84 78 109

CharizardMega
7 6 Fire Dragon 634 78 130 111 130
Charizard X

CharizardMega
8 6 Fire Flying 634 78 104 78 159 1
Charizard Y

In [17]:  poke.shape

Out[17]: (800, 13)

In [18]:  grass_pokemon.shape

Out[18]: (70, 13)

In [19]:  fire_pokemon.shape

Out[19]: (52, 13)

In [20]:  water_pokemon.shape

Out[20]: (112, 13)

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 4/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [21]:  # % of all 3 types

(112+70+52)/800

Out[21]: 0.2925

In [22]:  grass_pokemon.head()

Out[22]:
Sp.
# Name Primary_type Secondary_type Total HP Attack Defense
Atk

0 1 Bulbasaur Grass Poison 318 45 49 49 65

(https://getlin

1 2 Ivysaur Grass Poison 405 60 62 63 80

2 3 Venusaur Grass Poison 525 80 82 83 100

VenusaurMega
3 3 Grass Poison 625 80 100 123 122
Venusaur

48 43 Oddish Grass Poison 320 45 50 55 75

In [36]:  #to understand speed distribution

​
sns.histplot(grass_pokemon['Speed'])
plt.show()

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 5/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [37]:  sns.histplot(grass_pokemon['Sp. Atk'])

plt.show()
​
​

(https://getlin

In [38]:  sns.histplot(grass_pokemon['Sp. Def'])

plt.show()
​
​

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 6/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [30]:  grass_pokemon.describe()

Out[30]:
# Total HP Attack Defense Sp. Atk Sp.

count 70.000000 70.000000 70.000000 70.000000 70.000000 70.000000 70.000

mean 344.871429 421.142857 67.271429 73.214286 70.800000 77.500000 70.428

std 200.264385 106.650626 19.516564 25.380520 24.485192 27.244864 21.446

min 1.000000 180.000000 30.000000 27.000000 30.000000 24.000000 30.000

25% 187.250000 318.500000 51.250000 55.000000 50.000000 57.000000 55.000

(https://getlin
50% 372.000000 430.000000 65.500000 70.000000 66.000000 75.000000 66.000

75% 496.750000 497.000000 75.000000 93.500000 84.500000 99.500000 85.000

max 673.000000 630.000000 123.000000 132.000000 131.000000 145.000000 129.000

In [31]:  water_pokemon.head()

Out[31]:
Sp.
# Name Primary_type Secondary_type Total HP Attack Defense
Atk

9 7 Squirtle Water NaN 314 44 48 65 50

10 8 Wartortle Water NaN 405 59 63 80 65

11 9 Blastoise Water NaN 530 79 83 100 85

BlastoiseMega
12 9 Water NaN 630 79 103 120 135
Blastoise

59 54 Psyduck Water NaN 320 50 52 48 65

In [35]:  sns.histplot(water_pokemon['Speed'])
plt.show()
​

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 7/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [39]:  sns.histplot(water_pokemon['Sp. Atk'])

plt.show()

(https://getlin

In [40]:  sns.histplot(water_pokemon['Sp. Def'])

plt.show()

In [41]:  water_pokemon.describe()

Out[41]:
# Total HP Attack Defense Sp. Atk Sp.

count 112.000000 112.000000 112.000000 112.000000 112.000000 112.000000 112.000

mean 303.089286 430.455357 72.062500 74.151786 72.946429 74.812500 70.517

std 188.440807 113.188266 27.487026 28.377192 27.773809 29.030128 28.460

min 7.000000 200.000000 20.000000 10.000000 20.000000 10.000000 20.000

25% 130.000000 328.750000 52.250000 53.000000 54.500000 55.000000 50.000

50% 275.000000 455.000000 70.000000 72.000000 70.000000 70.000000 65.000

75% 456.250000 502.250000 90.250000 92.000000 88.500000 90.500000 89.250

max 693.000000 770.000000 170.000000 155.000000 180.000000 180.000000 160.000

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 8/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [42]:  grass_pokemon.describe()

Out[42]:
# Total HP Attack Defense Sp. Atk Sp.

count 70.000000 70.000000 70.000000 70.000000 70.000000 70.000000 70.000

mean 344.871429 421.142857 67.271429 73.214286 70.800000 77.500000 70.428

std 200.264385 106.650626 19.516564 25.380520 24.485192 27.244864 21.446

min 1.000000 180.000000 30.000000 27.000000 30.000000 24.000000 30.000

25% 187.250000 318.500000 51.250000 55.000000 50.000000 57.000000 55.000

(https://getlin
50% 372.000000 430.000000 65.500000 70.000000 66.000000 75.000000 66.000

75% 496.750000 497.000000 75.000000 93.500000 84.500000 99.500000 85.000

max 673.000000 630.000000 123.000000 132.000000 131.000000 145.000000 129.000

In [43]:  sns.histplot(fire_pokemon['Speed'], color='Orange')

plt.show()

In [44]:  sns.histplot(fire_pokemon['Sp. Atk'], color='Orange')

plt.show()

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 9/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [45]:  sns.histplot(fire_pokemon['Sp. Def'], color='Orange')

plt.show()

(https://getlin

In [46]:  fire_pokemon.describe()

Out[46]:
# Total HP Attack Defense Sp. Atk Sp.

count 52.000000 52.000000 52.000000 52.000000 52.000000 52.000000 52.000

mean 327.403846 458.076923 69.903846 84.769231 67.769231 88.980769 72.211

std 226.262840 109.760496 19.404123 28.769275 23.658200 30.042121 22.619

min 4.000000 250.000000 38.000000 30.000000 37.000000 15.000000 40.000

25% 143.500000 368.000000 58.000000 62.250000 51.000000 70.000000 54.750

50% 289.500000 482.000000 70.000000 84.500000 64.000000 85.000000 67.500

75% 513.250000 534.000000 80.000000 101.000000 78.000000 109.000000 85.000

max 721.000000 680.000000 115.000000 160.000000 140.000000 159.000000 154.000

In [47]:  water_pokemon.describe()

Out[47]:
# Total HP Attack Defense Sp. Atk Sp.

count 112.000000 112.000000 112.000000 112.000000 112.000000 112.000000 112.000

mean 303.089286 430.455357 72.062500 74.151786 72.946429 74.812500 70.517

std 188.440807 113.188266 27.487026 28.377192 27.773809 29.030128 28.460

min 7.000000 200.000000 20.000000 10.000000 20.000000 10.000000 20.000

25% 130.000000 328.750000 52.250000 53.000000 54.500000 55.000000 50.000

50% 275.000000 455.000000 70.000000 72.000000 70.000000 70.000000 65.000

75% 456.250000 502.250000 90.250000 92.000000 88.500000 90.500000 89.250

max 693.000000 770.000000 170.000000 155.000000 180.000000 180.000000 160.000

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 10/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [48]:  #sp def, sp attack and speed both higher in fire pokemon. mean also hig
​
#water pokemon loses from fire pokemon
​
#till now data analysis was done
​
​
​

In [49]:  #now apply decision tree algorithm

​
(https://getlin
poke.head()

Out[49]:
Sp. S
# Name Primary_type Secondary_type Total HP Attack Defense
Atk D

0 1 Bulbasaur Grass Poison 318 45 49 49 65 6

1 2 Ivysaur Grass Poison 405 60 62 63 80 8

2 3 Venusaur Grass Poison 525 80 82 83 100 10

VenusaurMega
3 3 Grass Poison 625 80 100 123 122 12
Venusaur

4 4 Charmander Fire NaN 309 39 52 43 60 5

In [50]:  poke['Legendary'].value_counts()

Out[50]: False 735

True 65
Name: Legendary, dtype: int64

In [51]:  #if we are comparing 2 features , presense of one fetaure is more than
#non legandary pokemons presence is much more than legandary
​
65/800

Out[51]: 0.08125

In [52]:  # 8 % presence of legandary pokemons

​
#to know relationship between speed and legandary column
​
#means classifying this pokemon is legandary or not on the basis of the
#so here we are using decision tree classifier
​
​
#so first divide dependent(legandary) variable and independent variable
​
​
​

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 11/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [54]:  x = poke[['Speed']]
y = poke[['Legendary']]

In [55]:  #divide data in traning and testing set and predict on test set
​
#for train and test set use sklearn
​
from sklearn.model_selection import train_test_split
​
x_train, x_test,y_train, y_test = train_test_split(x,y,test_size=0.3)
#means 30 % observations in test set rest 70 % training set observat
(https://getlin
​
​

In [56]:  #to make decison tree classifier

​
from sklearn.tree import DecisionTreeClassifier
#make instance
dtc = DecisionTreeClassifier()

In [57]:  #fit on train set

dtc.fit(x_train, y_train)

Out[57]: DecisionTreeClassifier()

In [58]:  #to predict value

​
y_pred = dtc.predict(x_test)

In [61]:  # to see difference in actual and predicted results

#to find accuracy import confusion matrix
​
from sklearn.metrics import confusion_matrix
confusion_matrix(y_test, y_pred)

Out[61]: array([[224, 1],

[ 15, 0]], dtype=int64)

In [62]:  #finding accuracy

​
(217+0)/(217+0+22+1)

Out[62]: 0.9041666666666667

In [63]:  #left diagonal shows correctly classified value and right diagonal show
#so i divide correctly classified values with all classified value so 9
#so implementation of decision tree done
​

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 12/13

03/11/2023, 09:27 decision tree on classification lab ML - Jupyter Notebook

In [ ]:  ​

(https://getlin

localhost:8888/notebooks/decision tree on classification lab ML.ipynb 13/13