{
"cells": [
{
"cell_type": "markdown",
"id": "7e9349d0",
"metadata": {},
"source": [
"# Poisoned Mushroom Dataset"
]
},
{
"cell_type": "markdown",
"id": "382f85ee",
"metadata": {},
"source": [
"We are going to take a quick tour of machine learning by working on an example dataset. The mushroom dataset\n",
"categorizes mushrooms as 'poisonous' or 'edible' and collects several descriptive properties of each mushroom example."
]
},
{
"cell_type": "code",
"execution_count": 41,
"id": "771ea7b9-e7de-43c7-96ae-bc3a30862715",
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd\n",
"import os"
]
},
{
"cell_type": "markdown",
"id": "c1121f80",
"metadata": {},
"source": [
"## Loading the dataset"
]
},
{
"cell_type": "code",
"execution_count": 118,
"id": "95d691a5-1638-4953-b929-761a8095a773",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
class
\n",
"
cap-shape
\n",
"
cap-surface
\n",
"
cap-color
\n",
"
ruises
\n",
"
odor
\n",
"
gill-attachment
\n",
"
gill-spacing
\n",
"
gill-size
\n",
"
gill-color
\n",
"
...
\n",
"
stalk-surface-below-ring
\n",
"
stalk-color-above-ring
\n",
"
stalk-color-below-ring
\n",
"
veil-type
\n",
"
veil-color
\n",
"
ring-number
\n",
"
ring-type
\n",
"
spore-print-color
\n",
"
population
\n",
"
habitat
\n",
"
\n",
" \n",
" \n",
"
\n",
"
0
\n",
"
e
\n",
"
x
\n",
"
f
\n",
"
n
\n",
"
f
\n",
"
n
\n",
"
f
\n",
"
w
\n",
"
n
\n",
"
b
\n",
"
...
\n",
"
y
\n",
"
w
\n",
"
p
\n",
"
NaN
\n",
"
n
\n",
"
o
\n",
"
p
\n",
"
w
\n",
"
v
\n",
"
NaN
\n",
"
\n",
"
\n",
"
1
\n",
"
p
\n",
"
NaN
\n",
"
y
\n",
"
g
\n",
"
t
\n",
"
NaN
\n",
"
f
\n",
"
c
\n",
"
b
\n",
"
k
\n",
"
...
\n",
"
s
\n",
"
n
\n",
"
c
\n",
"
p
\n",
"
w
\n",
"
n
\n",
"
e
\n",
"
NaN
\n",
"
y
\n",
"
g
\n",
"
\n",
"
\n",
"
2
\n",
"
e
\n",
"
b
\n",
"
y
\n",
"
n
\n",
"
t
\n",
"
n
\n",
"
f
\n",
"
c
\n",
"
NaN
\n",
"
n
\n",
"
...
\n",
"
s
\n",
"
p
\n",
"
NaN
\n",
"
p
\n",
"
w
\n",
"
o
\n",
"
p
\n",
"
b
\n",
"
y
\n",
"
w
\n",
"
\n",
"
\n",
"
3
\n",
"
e
\n",
"
x
\n",
"
g
\n",
"
g
\n",
"
t
\n",
"
n
\n",
"
f
\n",
"
w
\n",
"
b
\n",
"
n
\n",
"
...
\n",
"
s
\n",
"
p
\n",
"
NaN
\n",
"
p
\n",
"
w
\n",
"
n
\n",
"
n
\n",
"
NaN
\n",
"
NaN
\n",
"
d
\n",
"
\n",
"
\n",
"
4
\n",
"
e
\n",
"
NaN
\n",
"
f
\n",
"
NaN
\n",
"
t
\n",
"
n
\n",
"
a
\n",
"
w
\n",
"
n
\n",
"
n
\n",
"
...
\n",
"
k
\n",
"
NaN
\n",
"
w
\n",
"
p
\n",
"
w
\n",
"
NaN
\n",
"
l
\n",
"
w
\n",
"
v
\n",
"
d
\n",
"
\n",
"
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
\n",
"
\n",
"
25981
\n",
"
e
\n",
"
f
\n",
"
NaN
\n",
"
r
\n",
"
f
\n",
"
n
\n",
"
f
\n",
"
NaN
\n",
"
n
\n",
"
NaN
\n",
"
...
\n",
"
NaN
\n",
"
n
\n",
"
p
\n",
"
p
\n",
"
w
\n",
"
o
\n",
"
p
\n",
"
k
\n",
"
v
\n",
"
NaN
\n",
"
\n",
"
\n",
"
25982
\n",
"
e
\n",
"
f
\n",
"
s
\n",
"
e
\n",
"
f
\n",
"
NaN
\n",
"
f
\n",
"
c
\n",
"
n
\n",
"
y
\n",
"
...
\n",
"
y
\n",
"
w
\n",
"
p
\n",
"
p
\n",
"
w
\n",
"
NaN
\n",
"
p
\n",
"
r
\n",
"
y
\n",
"
d
\n",
"
\n",
"
\n",
"
25983
\n",
"
p
\n",
"
f
\n",
"
g
\n",
"
e
\n",
"
NaN
\n",
"
NaN
\n",
"
a
\n",
"
c
\n",
"
b
\n",
"
b
\n",
"
...
\n",
"
y
\n",
"
w
\n",
"
NaN
\n",
"
p
\n",
"
w
\n",
"
o
\n",
"
p
\n",
"
h
\n",
"
v
\n",
"
m
\n",
"
\n",
"
\n",
"
25984
\n",
"
e
\n",
"
x
\n",
"
g
\n",
"
g
\n",
"
t
\n",
"
n
\n",
"
f
\n",
"
w
\n",
"
b
\n",
"
h
\n",
"
...
\n",
"
f
\n",
"
NaN
\n",
"
NaN
\n",
"
p
\n",
"
w
\n",
"
t
\n",
"
e
\n",
"
NaN
\n",
"
s
\n",
"
NaN
\n",
"
\n",
"
\n",
"
25985
\n",
"
e
\n",
"
b
\n",
"
y
\n",
"
y
\n",
"
t
\n",
"
l
\n",
"
f
\n",
"
c
\n",
"
b
\n",
"
y
\n",
"
...
\n",
"
k
\n",
"
g
\n",
"
o
\n",
"
p
\n",
"
w
\n",
"
o
\n",
"
l
\n",
"
k
\n",
"
s
\n",
"
g
\n",
"
\n",
" \n",
"
\n",
"
25986 rows × 23 columns
\n",
"
"
],
"text/plain": [
" class cap-shape cap-surface cap-color ruises odor gill-attachment \\\n",
"0 e x f n f n f \n",
"1 p NaN y g t NaN f \n",
"2 e b y n t n f \n",
"3 e x g g t n f \n",
"4 e NaN f NaN t n a \n",
"... ... ... ... ... ... ... ... \n",
"25981 e f NaN r f n f \n",
"25982 e f s e f NaN f \n",
"25983 p f g e NaN NaN a \n",
"25984 e x g g t n f \n",
"25985 e b y y t l f \n",
"\n",
" gill-spacing gill-size gill-color ... stalk-surface-below-ring \\\n",
"0 w n b ... y \n",
"1 c b k ... s \n",
"2 c NaN n ... s \n",
"3 w b n ... s \n",
"4 w n n ... k \n",
"... ... ... ... ... ... \n",
"25981 NaN n NaN ... NaN \n",
"25982 c n y ... y \n",
"25983 c b b ... y \n",
"25984 w b h ... f \n",
"25985 c b y ... k \n",
"\n",
" stalk-color-above-ring stalk-color-below-ring veil-type veil-color \\\n",
"0 w p NaN n \n",
"1 n c p w \n",
"2 p NaN p w \n",
"3 p NaN p w \n",
"4 NaN w p w \n",
"... ... ... ... ... \n",
"25981 n p p w \n",
"25982 w p p w \n",
"25983 w NaN p w \n",
"25984 NaN NaN p w \n",
"25985 g o p w \n",
"\n",
" ring-number ring-type spore-print-color population habitat \n",
"0 o p w v NaN \n",
"1 n e NaN y g \n",
"2 o p b y w \n",
"3 n n NaN NaN d \n",
"4 NaN l w v d \n",
"... ... ... ... ... ... \n",
"25981 o p k v NaN \n",
"25982 NaN p r y d \n",
"25983 o p h v m \n",
"25984 t e NaN s NaN \n",
"25985 o l k s g \n",
"\n",
"[25986 rows x 23 columns]"
]
},
"execution_count": 118,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# These lines would load the data locally\n",
"# data_root = \"./\"\n",
"# filename = \"mushroom.csv\"\n",
"# filepath = os.path.join(data_root, filename)\n",
"\n",
"# We'll fetch it directly from the web\n",
"data_url = \"https://aet-cs.github.io/white/ML/lessons/mushroom.csv\"\n",
"df = pd.read_csv(data_url)\n",
"df"
]
},
{
"cell_type": "markdown",
"id": "a71ed4f4-5106-4425-88cd-ab2acbc24b86",
"metadata": {},
"source": [
"`describe` gives a quick overview of each feature"
]
},
{
"cell_type": "code",
"execution_count": 43,
"id": "a363da42-73b6-4d09-b1c6-f885b020c633",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
class
\n",
"
cap-shape
\n",
"
cap-surface
\n",
"
cap-color
\n",
"
ruises
\n",
"
odor
\n",
"
gill-attachment
\n",
"
gill-spacing
\n",
"
gill-size
\n",
"
gill-color
\n",
"
...
\n",
"
stalk-surface-below-ring
\n",
"
stalk-color-above-ring
\n",
"
stalk-color-below-ring
\n",
"
veil-type
\n",
"
veil-color
\n",
"
ring-number
\n",
"
ring-type
\n",
"
spore-print-color
\n",
"
population
\n",
"
habitat
\n",
"
\n",
" \n",
" \n",
"
\n",
"
count
\n",
"
25986
\n",
"
22513
\n",
"
22507
\n",
"
22527
\n",
"
22514
\n",
"
22536
\n",
"
22505
\n",
"
22587
\n",
"
22494
\n",
"
22418
\n",
"
...
\n",
"
22563
\n",
"
22413
\n",
"
22553
\n",
"
22489
\n",
"
22483
\n",
"
22497
\n",
"
22478
\n",
"
22493
\n",
"
22475
\n",
"
22502
\n",
"
\n",
"
\n",
"
unique
\n",
"
2
\n",
"
6
\n",
"
4
\n",
"
10
\n",
"
2
\n",
"
9
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
12
\n",
"
...
\n",
"
4
\n",
"
9
\n",
"
9
\n",
"
1
\n",
"
4
\n",
"
3
\n",
"
5
\n",
"
9
\n",
"
6
\n",
"
7
\n",
"
\n",
"
\n",
"
top
\n",
"
e
\n",
"
x
\n",
"
y
\n",
"
n
\n",
"
f
\n",
"
n
\n",
"
f
\n",
"
c
\n",
"
b
\n",
"
b
\n",
"
...
\n",
"
s
\n",
"
w
\n",
"
w
\n",
"
p
\n",
"
w
\n",
"
o
\n",
"
p
\n",
"
w
\n",
"
v
\n",
"
d
\n",
"
\n",
"
\n",
"
freq
\n",
"
14354
\n",
"
7674
\n",
"
7602
\n",
"
4810
\n",
"
12361
\n",
"
6986
\n",
"
17811
\n",
"
16092
\n",
"
13997
\n",
"
3679
\n",
"
...
\n",
"
10619
\n",
"
8580
\n",
"
8403
\n",
"
22489
\n",
"
15742
\n",
"
15713
\n",
"
8501
\n",
"
5085
\n",
"
8409
\n",
"
6573
\n",
"
\n",
" \n",
"
\n",
"
4 rows × 23 columns
\n",
"
"
],
"text/plain": [
" class cap-shape cap-surface cap-color ruises odor gill-attachment \\\n",
"count 25986 22513 22507 22527 22514 22536 22505 \n",
"unique 2 6 4 10 2 9 2 \n",
"top e x y n f n f \n",
"freq 14354 7674 7602 4810 12361 6986 17811 \n",
"\n",
" gill-spacing gill-size gill-color ... stalk-surface-below-ring \\\n",
"count 22587 22494 22418 ... 22563 \n",
"unique 2 2 12 ... 4 \n",
"top c b b ... s \n",
"freq 16092 13997 3679 ... 10619 \n",
"\n",
" stalk-color-above-ring stalk-color-below-ring veil-type veil-color \\\n",
"count 22413 22553 22489 22483 \n",
"unique 9 9 1 4 \n",
"top w w p w \n",
"freq 8580 8403 22489 15742 \n",
"\n",
" ring-number ring-type spore-print-color population habitat \n",
"count 22497 22478 22493 22475 22502 \n",
"unique 3 5 9 6 7 \n",
"top o p w v d \n",
"freq 15713 8501 5085 8409 6573 \n",
"\n",
"[4 rows x 23 columns]"
]
},
"execution_count": 43,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.describe()"
]
},
{
"cell_type": "markdown",
"id": "dc3a4445-274a-4a7f-b397-6f96ddc26d7a",
"metadata": {},
"source": [
"All the features in this dataset are categorical, with between 2 and 12 categories. Except the curious `veil-type` which has only one value. Since `veil-type` has only one unique value, we'll drop it."
]
},
{
"cell_type": "code",
"execution_count": 119,
"id": "9494a6c7-67e8-419c-9aaa-0f8ff89221bb",
"metadata": {},
"outputs": [],
"source": [
"df = df.drop('veil-type', axis=1)"
]
},
{
"cell_type": "markdown",
"id": "b7c00c5e",
"metadata": {},
"source": [
"## Data Exploration"
]
},
{
"cell_type": "markdown",
"id": "66b733f8-c4db-41d1-8e9e-ac316c0b8ecb",
"metadata": {},
"source": [
"Show all the columns. Notice the target is the first column!"
]
},
{
"cell_type": "code",
"execution_count": 120,
"id": "91eae28c-7318-4630-b527-3cb6763babeb",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Index(['class', 'cap-shape', 'cap-surface', 'cap-color', 'ruises', 'odor',\n",
" 'gill-attachment', 'gill-spacing', 'gill-size', 'gill-color',\n",
" 'stalk-shape', 'stalk-root', 'stalk-surface-above-ring',\n",
" 'stalk-surface-below-ring', 'stalk-color-above-ring',\n",
" 'stalk-color-below-ring', 'veil-color', 'ring-number', 'ring-type',\n",
" 'spore-print-color', 'population', 'habitat'],\n",
" dtype='object')"
]
},
"execution_count": 120,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.columns"
]
},
{
"cell_type": "markdown",
"id": "b5a2457e-739c-41ca-a91e-bbabbd5686f5",
"metadata": {},
"source": [
"Get the size of the dataframe. Shape returns (rows, cols)"
]
},
{
"cell_type": "code",
"execution_count": 121,
"id": "7b8d5fcc-1043-427a-9264-6f0a5deeb995",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(25986, 22)"
]
},
"execution_count": 121,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.shape"
]
},
{
"cell_type": "markdown",
"id": "3e22afb6-b203-4542-9c36-f4910be75a60",
"metadata": {},
"source": [
"Let's peek at the target"
]
},
{
"cell_type": "code",
"execution_count": 122,
"id": "2c1e74a9-08b9-4e73-8855-8af51df9a2ea",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"0 e\n",
"1 p\n",
"2 e\n",
"3 e\n",
"4 e\n",
" ..\n",
"25981 e\n",
"25982 e\n",
"25983 p\n",
"25984 e\n",
"25985 e\n",
"Name: class, Length: 25986, dtype: object"
]
},
"execution_count": 122,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df['class']"
]
},
{
"cell_type": "markdown",
"id": "a2d5b6d9-e9dc-4bc6-808f-a5a5523f585c",
"metadata": {},
"source": [
"This dataset has a LOT of \"N/A\" datapoints. One way to clean the data is to drop all affected rows"
]
},
{
"cell_type": "code",
"execution_count": 123,
"id": "635ab8ad-939b-4451-ae75-7414cc9436f1",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(1327, 22)"
]
},
"execution_count": 123,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.dropna().shape"
]
},
{
"cell_type": "markdown",
"id": "13baf05c-9ed1-4d7f-a794-3f8d6ae332d5",
"metadata": {},
"source": [
"But this *significantly* reduces our dataset. Let's instead use a data imputation strategy that fills the N/A with the *mode*"
]
},
{
"cell_type": "code",
"execution_count": 124,
"id": "90fc05fd-0f98-4d68-bce8-4bf9c3cd77db",
"metadata": {},
"outputs": [],
"source": [
"for c in df.columns:\n",
" df = df.fillna({c: df[c].mode()[0]})"
]
},
{
"cell_type": "markdown",
"id": "5aebfa3d-f8c3-4da8-a28f-fdc8bbe6f40f",
"metadata": {},
"source": [
"Look at df again"
]
},
{
"cell_type": "code",
"execution_count": 125,
"id": "63191ab4-1b33-49fe-b9c3-8a9cfce68f7d",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
class
\n",
"
cap-shape
\n",
"
cap-surface
\n",
"
cap-color
\n",
"
ruises
\n",
"
odor
\n",
"
gill-attachment
\n",
"
gill-spacing
\n",
"
gill-size
\n",
"
gill-color
\n",
"
...
\n",
"
stalk-surface-above-ring
\n",
"
stalk-surface-below-ring
\n",
"
stalk-color-above-ring
\n",
"
stalk-color-below-ring
\n",
"
veil-color
\n",
"
ring-number
\n",
"
ring-type
\n",
"
spore-print-color
\n",
"
population
\n",
"
habitat
\n",
"
\n",
" \n",
" \n",
"
\n",
"
0
\n",
"
e
\n",
"
x
\n",
"
f
\n",
"
n
\n",
"
f
\n",
"
n
\n",
"
f
\n",
"
w
\n",
"
n
\n",
"
b
\n",
"
...
\n",
"
s
\n",
"
y
\n",
"
w
\n",
"
p
\n",
"
n
\n",
"
o
\n",
"
p
\n",
"
w
\n",
"
v
\n",
"
d
\n",
"
\n",
"
\n",
"
1
\n",
"
p
\n",
"
x
\n",
"
y
\n",
"
g
\n",
"
t
\n",
"
n
\n",
"
f
\n",
"
c
\n",
"
b
\n",
"
k
\n",
"
...
\n",
"
f
\n",
"
s
\n",
"
n
\n",
"
c
\n",
"
w
\n",
"
n
\n",
"
e
\n",
"
w
\n",
"
y
\n",
"
g
\n",
"
\n",
"
\n",
"
2
\n",
"
e
\n",
"
b
\n",
"
y
\n",
"
n
\n",
"
t
\n",
"
n
\n",
"
f
\n",
"
c
\n",
"
b
\n",
"
n
\n",
"
...
\n",
"
s
\n",
"
s
\n",
"
p
\n",
"
w
\n",
"
w
\n",
"
o
\n",
"
p
\n",
"
b
\n",
"
y
\n",
"
w
\n",
"
\n",
"
\n",
"
3
\n",
"
e
\n",
"
x
\n",
"
g
\n",
"
g
\n",
"
t
\n",
"
n
\n",
"
f
\n",
"
w
\n",
"
b
\n",
"
n
\n",
"
...
\n",
"
s
\n",
"
s
\n",
"
p
\n",
"
w
\n",
"
w
\n",
"
n
\n",
"
n
\n",
"
w
\n",
"
v
\n",
"
d
\n",
"
\n",
"
\n",
"
4
\n",
"
e
\n",
"
x
\n",
"
f
\n",
"
n
\n",
"
t
\n",
"
n
\n",
"
a
\n",
"
w
\n",
"
n
\n",
"
n
\n",
"
...
\n",
"
s
\n",
"
k
\n",
"
w
\n",
"
w
\n",
"
w
\n",
"
o
\n",
"
l
\n",
"
w
\n",
"
v
\n",
"
d
\n",
"
\n",
"
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
...
\n",
"
\n",
"
\n",
"
25981
\n",
"
e
\n",
"
f
\n",
"
y
\n",
"
r
\n",
"
f
\n",
"
n
\n",
"
f
\n",
"
c
\n",
"
n
\n",
"
b
\n",
"
...
\n",
"
s
\n",
"
s
\n",
"
n
\n",
"
p
\n",
"
w
\n",
"
o
\n",
"
p
\n",
"
k
\n",
"
v
\n",
"
d
\n",
"
\n",
"
\n",
"
25982
\n",
"
e
\n",
"
f
\n",
"
s
\n",
"
e
\n",
"
f
\n",
"
n
\n",
"
f
\n",
"
c
\n",
"
n
\n",
"
y
\n",
"
...
\n",
"
s
\n",
"
y
\n",
"
w
\n",
"
p
\n",
"
w
\n",
"
o
\n",
"
p
\n",
"
r
\n",
"
y
\n",
"
d
\n",
"
\n",
"
\n",
"
25983
\n",
"
p
\n",
"
f
\n",
"
g
\n",
"
e
\n",
"
f
\n",
"
n
\n",
"
a
\n",
"
c
\n",
"
b
\n",
"
b
\n",
"
...
\n",
"
s
\n",
"
y
\n",
"
w
\n",
"
w
\n",
"
w
\n",
"
o
\n",
"
p
\n",
"
h
\n",
"
v
\n",
"
m
\n",
"
\n",
"
\n",
"
25984
\n",
"
e
\n",
"
x
\n",
"
g
\n",
"
g
\n",
"
t
\n",
"
n
\n",
"
f
\n",
"
w
\n",
"
b
\n",
"
h
\n",
"
...
\n",
"
k
\n",
"
f
\n",
"
w
\n",
"
w
\n",
"
w
\n",
"
t
\n",
"
e
\n",
"
w
\n",
"
s
\n",
"
d
\n",
"
\n",
"
\n",
"
25985
\n",
"
e
\n",
"
b
\n",
"
y
\n",
"
y
\n",
"
t
\n",
"
l
\n",
"
f
\n",
"
c
\n",
"
b
\n",
"
y
\n",
"
...
\n",
"
k
\n",
"
k
\n",
"
g
\n",
"
o
\n",
"
w
\n",
"
o
\n",
"
l
\n",
"
k
\n",
"
s
\n",
"
g
\n",
"
\n",
" \n",
"
\n",
"
25986 rows × 22 columns
\n",
"
"
],
"text/plain": [
" class cap-shape cap-surface cap-color ruises odor gill-attachment \\\n",
"0 e x f n f n f \n",
"1 p x y g t n f \n",
"2 e b y n t n f \n",
"3 e x g g t n f \n",
"4 e x f n t n a \n",
"... ... ... ... ... ... ... ... \n",
"25981 e f y r f n f \n",
"25982 e f s e f n f \n",
"25983 p f g e f n a \n",
"25984 e x g g t n f \n",
"25985 e b y y t l f \n",
"\n",
" gill-spacing gill-size gill-color ... stalk-surface-above-ring \\\n",
"0 w n b ... s \n",
"1 c b k ... f \n",
"2 c b n ... s \n",
"3 w b n ... s \n",
"4 w n n ... s \n",
"... ... ... ... ... ... \n",
"25981 c n b ... s \n",
"25982 c n y ... s \n",
"25983 c b b ... s \n",
"25984 w b h ... k \n",
"25985 c b y ... k \n",
"\n",
" stalk-surface-below-ring stalk-color-above-ring stalk-color-below-ring \\\n",
"0 y w p \n",
"1 s n c \n",
"2 s p w \n",
"3 s p w \n",
"4 k w w \n",
"... ... ... ... \n",
"25981 s n p \n",
"25982 y w p \n",
"25983 y w w \n",
"25984 f w w \n",
"25985 k g o \n",
"\n",
" veil-color ring-number ring-type spore-print-color population habitat \n",
"0 n o p w v d \n",
"1 w n e w y g \n",
"2 w o p b y w \n",
"3 w n n w v d \n",
"4 w o l w v d \n",
"... ... ... ... ... ... ... \n",
"25981 w o p k v d \n",
"25982 w o p r y d \n",
"25983 w o p h v m \n",
"25984 w t e w s d \n",
"25985 w o l k s g \n",
"\n",
"[25986 rows x 22 columns]"
]
},
"execution_count": 125,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df"
]
},
{
"cell_type": "markdown",
"id": "df21f38d-9f68-46ca-bf03-682cd3e49315",
"metadata": {},
"source": [
"Let's see what the classifications are and how balanced the dataset is. Highly imbalanced datasets require special techniques to ensure valid models."
]
},
{
"cell_type": "code",
"execution_count": 127,
"id": "e7f294dd-8f1e-47dd-9a12-9656ba1e893f",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"class\n",
"e 14354\n",
"p 11632\n",
"Name: count, dtype: int64"
]
},
"execution_count": 127,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df['class'].value_counts()"
]
},
{
"cell_type": "markdown",
"id": "33d056df-7b9b-41db-8760-f1dec051f951",
"metadata": {},
"source": [
"We'll introduce a new plotting library -- \"seaborn\", which has some advantages over matplotlib. Here we show how to quickly make a histogram from a dataframe. Seaborn works nicely with pandas dataframes."
]
},
{
"cell_type": "code",
"execution_count": 128,
"id": "fd3c5ce9-dbfb-4bed-8105-8587a23c35c6",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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",
"text/plain": [
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# Count plot\n",
"sns.countplot(x='cap-color', data=df, )\n",
"plt.title('Count Plot of Cap Color Frequencies')\n",
"plt.xlabel('Cap Color')\n",
"plt.ylabel('Frequency')\n",
"plt.show()\n"
]
},
{
"cell_type": "markdown",
"id": "7a2f82fe-774f-4a74-98f8-8a14a4a5e343",
"metadata": {},
"source": [
"I wonder how the color correlates to the outcome -- are some color more poisonous? We'll do some pandas work to make this summary for us. (Here's a nice overview of `groupby`: https://builtin.com/data-science/pandas-groupby)"
]
},
{
"cell_type": "code",
"execution_count": 175,
"id": "618f14e5",
"metadata": {},
"outputs": [],
"source": [
"# Count observations by color and toxicity\n",
"counts = df.groupby(['cap-color', 'class']).size().reset_index(name='count')"
]
},
{
"cell_type": "code",
"execution_count": 176,
"id": "f0d0b10f-85a4-4f4d-9c5f-7d1dc6a93dd0",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
cap-color
\n",
"
class
\n",
"
count
\n",
"
\n",
" \n",
" \n",
"
\n",
"
0
\n",
"
b
\n",
"
e
\n",
"
660
\n",
"
\n",
"
\n",
"
1
\n",
"
b
\n",
"
p
\n",
"
551
\n",
"
\n",
"
\n",
"
2
\n",
"
c
\n",
"
e
\n",
"
526
\n",
"
\n",
"
\n",
"
3
\n",
"
c
\n",
"
p
\n",
"
413
\n",
"
\n",
"
\n",
"
4
\n",
"
e
\n",
"
e
\n",
"
1834
\n",
"
\n",
"
\n",
"
5
\n",
"
e
\n",
"
p
\n",
"
1627
\n",
"
\n",
"
\n",
"
6
\n",
"
g
\n",
"
e
\n",
"
2226
\n",
"
\n",
"
\n",
"
7
\n",
"
g
\n",
"
p
\n",
"
1765
\n",
"
\n",
"
\n",
"
8
\n",
"
n
\n",
"
e
\n",
"
2673
\n",
"
\n",
"
\n",
"
9
\n",
"
n
\n",
"
p
\n",
"
2137
\n",
"
\n",
"
\n",
"
10
\n",
"
p
\n",
"
e
\n",
"
611
\n",
"
\n",
"
\n",
"
11
\n",
"
p
\n",
"
p
\n",
"
483
\n",
"
\n",
"
\n",
"
12
\n",
"
r
\n",
"
e
\n",
"
467
\n",
"
\n",
"
\n",
"
13
\n",
"
r
\n",
"
p
\n",
"
380
\n",
"
\n",
"
\n",
"
14
\n",
"
u
\n",
"
e
\n",
"
502
\n",
"
\n",
"
\n",
"
15
\n",
"
u
\n",
"
p
\n",
"
413
\n",
"
\n",
"
\n",
"
16
\n",
"
w
\n",
"
e
\n",
"
1452
\n",
"
\n",
"
\n",
"
17
\n",
"
w
\n",
"
p
\n",
"
1137
\n",
"
\n",
"
\n",
"
18
\n",
"
y
\n",
"
e
\n",
"
1460
\n",
"
\n",
"
\n",
"
19
\n",
"
y
\n",
"
p
\n",
"
1210
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" cap-color class count\n",
"0 b e 660\n",
"1 b p 551\n",
"2 c e 526\n",
"3 c p 413\n",
"4 e e 1834\n",
"5 e p 1627\n",
"6 g e 2226\n",
"7 g p 1765\n",
"8 n e 2673\n",
"9 n p 2137\n",
"10 p e 611\n",
"11 p p 483\n",
"12 r e 467\n",
"13 r p 380\n",
"14 u e 502\n",
"15 u p 413\n",
"16 w e 1452\n",
"17 w p 1137\n",
"18 y e 1460\n",
"19 y p 1210"
]
},
"execution_count": 176,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"counts"
]
},
{
"cell_type": "code",
"execution_count": 165,
"id": "931ce993-ecce-4aee-98b6-df53b08a66f6",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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",
"text/plain": [
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# Create the bar plot\n",
"plt.figure(figsize=(10, 6))\n",
"sns.barplot(x='cap-color', y='count', hue='class', data=counts, palette={'p': 'blue', 'e': 'red'})\n",
"\n",
"# Add plot title and labels\n",
"plt.title('Distribution of Mushroom Colors with Poisonous Indication')\n",
"plt.xlabel('Color')\n",
"plt.ylabel('Count')\n",
"\n",
"\n",
"# Customize the legend\n",
"legend = plt.legend(title='Toxicity', labels=['Edible', 'Poisonous'])\n",
"legend.get_texts()[0].set_color('red') # Edible in red\n",
"legend.get_texts()[1].set_color('blue') # Poisonous in blue\n",
"\n",
"\n",
"# Show the plot\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 143,
"id": "405b65e9-f10a-4d98-b812-4fadff77660f",
"metadata": {
"scrolled": true
},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
class
\n",
"
cap-shape
\n",
"
cap-surface
\n",
"
cap-color
\n",
"
ruises
\n",
"
odor
\n",
"
gill-attachment
\n",
"
gill-spacing
\n",
"
gill-size
\n",
"
gill-color
\n",
"
...
\n",
"
stalk-surface-above-ring
\n",
"
stalk-surface-below-ring
\n",
"
stalk-color-above-ring
\n",
"
stalk-color-below-ring
\n",
"
veil-color
\n",
"
ring-number
\n",
"
ring-type
\n",
"
spore-print-color
\n",
"
population
\n",
"
habitat
\n",
"
\n",
" \n",
" \n",
"
\n",
"
count
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
...
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
\n",
"
\n",
"
unique
\n",
"
2
\n",
"
6
\n",
"
4
\n",
"
10
\n",
"
2
\n",
"
9
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
12
\n",
"
...
\n",
"
4
\n",
"
4
\n",
"
9
\n",
"
9
\n",
"
4
\n",
"
3
\n",
"
5
\n",
"
9
\n",
"
6
\n",
"
7
\n",
"
\n",
"
\n",
"
top
\n",
"
e
\n",
"
x
\n",
"
y
\n",
"
n
\n",
"
f
\n",
"
n
\n",
"
f
\n",
"
c
\n",
"
b
\n",
"
b
\n",
"
...
\n",
"
s
\n",
"
s
\n",
"
w
\n",
"
w
\n",
"
w
\n",
"
o
\n",
"
p
\n",
"
w
\n",
"
v
\n",
"
d
\n",
"
\n",
"
\n",
"
freq
\n",
"
14354
\n",
"
11147
\n",
"
11081
\n",
"
8269
\n",
"
15833
\n",
"
10436
\n",
"
21292
\n",
"
19491
\n",
"
17489
\n",
"
7247
\n",
"
...
\n",
"
14548
\n",
"
14042
\n",
"
12153
\n",
"
11836
\n",
"
19245
\n",
"
19202
\n",
"
12009
\n",
"
8578
\n",
"
11920
\n",
"
10057
\n",
"
\n",
" \n",
"
\n",
"
4 rows × 22 columns
\n",
"
"
],
"text/plain": [
" class cap-shape cap-surface cap-color ruises odor gill-attachment \\\n",
"count 25986 25986 25986 25986 25986 25986 25986 \n",
"unique 2 6 4 10 2 9 2 \n",
"top e x y n f n f \n",
"freq 14354 11147 11081 8269 15833 10436 21292 \n",
"\n",
" gill-spacing gill-size gill-color ... stalk-surface-above-ring \\\n",
"count 25986 25986 25986 ... 25986 \n",
"unique 2 2 12 ... 4 \n",
"top c b b ... s \n",
"freq 19491 17489 7247 ... 14548 \n",
"\n",
" stalk-surface-below-ring stalk-color-above-ring stalk-color-below-ring \\\n",
"count 25986 25986 25986 \n",
"unique 4 9 9 \n",
"top s w w \n",
"freq 14042 12153 11836 \n",
"\n",
" veil-color ring-number ring-type spore-print-color population habitat \n",
"count 25986 25986 25986 25986 25986 25986 \n",
"unique 4 3 5 9 6 7 \n",
"top w o p w v d \n",
"freq 19245 19202 12009 8578 11920 10057 \n",
"\n",
"[4 rows x 22 columns]"
]
},
"execution_count": 143,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.describe()"
]
},
{
"cell_type": "code",
"execution_count": 144,
"id": "e8187689-c4d0-4181-aff4-666b0a43138b",
"metadata": {},
"outputs": [],
"source": [
"# Count observations by odor and toxicity\n",
"counts = df.groupby(['odor', 'class']).size().reset_index(name='count')"
]
},
{
"cell_type": "code",
"execution_count": 145,
"id": "2e3c3a6c-4d7b-4fed-befe-0ef7f74147fe",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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",
"text/plain": [
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# Create the bar plot\n",
"plt.figure(figsize=(10, 6))\n",
"sns.barplot(x='odor', y='count', hue='class', data=counts, palette={'p': 'blue', 'e': 'red'})\n",
"\n",
"# Add plot title and labels\n",
"plt.title('Distribution of Mushroom Odor with Poisonous Indication')\n",
"plt.xlabel('Odor')\n",
"plt.ylabel('Count')\n",
"\n",
"# Customize the legend\n",
"legend = plt.legend(title='Toxicity', labels=['Edible', 'Poisonous'])\n",
"legend.get_texts()[0].set_color('red') # Edible in red\n",
"legend.get_texts()[1].set_color('blue') # Poisonous in blue\n",
"\n",
"# Show the plot\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"id": "0bdfeadc-13f3-4041-a954-d46ed9ead5b2",
"metadata": {},
"source": [
"## Correlation matrix heat map"
]
},
{
"cell_type": "markdown",
"id": "7fc020d8-adfa-48de-b9bb-9dabde9499a4",
"metadata": {},
"source": [
"Let's get a quick visual representation of the relationshop between features in this dataset. We'll use a version of a Chi-Squared test on all pairs $(n,m)$ of features in the dataset, including the target. (Heat maps for continuous data are easy to plot -- because of the categories we have to do some extra work here. You can treat `cramers_v` as a black box for now.)"
]
},
{
"cell_type": "code",
"execution_count": 62,
"id": "1cd6565e-12d2-48a1-bfc1-35c44e39e22b",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# Plotting the correlation matrix\n",
"plt.figure(figsize=(10,8))\n",
"sns.heatmap(corr_matrix, annot=False, cmap='Blues', square=True, cbar_kws={\"shrink\": .8})\n",
"plt.title(\"Cramér's V Correlation Matrix for Categorical Features\")\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"id": "e406b72a-e7e7-4294-853c-66ec7f36f6bc",
"metadata": {},
"source": [
"Which features seem to be important?"
]
},
{
"cell_type": "code",
"execution_count": 64,
"id": "7ff1e04d-41dd-48b9-97e4-c17ca7351f75",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"class 0.999922\n",
"odor 0.465191\n",
"gill-size 0.122908\n",
"stalk-surface-above-ring 0.129749\n",
"stalk-surface-below-ring 0.102529\n",
"spore-print-color 0.124672\n",
"Name: class, dtype: float64"
]
},
"execution_count": 64,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"corr_matrix['class'][corr_matrix['class']>0.1]"
]
},
{
"cell_type": "markdown",
"id": "961a4b15-dbfb-4585-86f6-5d0bddd1f444",
"metadata": {},
"source": [
"## Data Modeling"
]
},
{
"cell_type": "markdown",
"id": "7416c6a5-a15e-4fc9-81a1-529ebdd44b6b",
"metadata": {},
"source": [
"We're finally ready to do some data modeling using scikit-learn. In this cell we import some methods we'll use, reload the data frame (just to be safe), re-pre-process-it, and one-hot-encode all the categorical variables."
]
},
{
"cell_type": "code",
"execution_count": 148,
"id": "f5cf2242-b167-4111-a509-ede1e11c8c75",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
cap-shape_b
\n",
"
cap-shape_c
\n",
"
cap-shape_f
\n",
"
cap-shape_k
\n",
"
cap-shape_s
\n",
"
cap-shape_x
\n",
"
cap-surface_f
\n",
"
cap-surface_g
\n",
"
cap-surface_s
\n",
"
cap-surface_y
\n",
"
...
\n",
"
population_s
\n",
"
population_v
\n",
"
population_y
\n",
"
habitat_d
\n",
"
habitat_g
\n",
"
habitat_l
\n",
"
habitat_m
\n",
"
habitat_p
\n",
"
habitat_u
\n",
"
habitat_w
\n",
"
\n",
" \n",
" \n",
"
\n",
"
count
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
...
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
25986
\n",
"
\n",
"
\n",
"
unique
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
...
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
2
\n",
"
\n",
"
\n",
"
top
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
...
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
False
\n",
"
\n",
"
\n",
"
freq
\n",
"
23726
\n",
"
24509
\n",
"
19264
\n",
"
23074
\n",
"
24518
\n",
"
18312
\n",
"
19762
\n",
"
23825
\n",
"
19466
\n",
"
18384
\n",
"
...
\n",
"
22365
\n",
"
17577
\n",
"
21675
\n",
"
19413
\n",
"
21040
\n",
"
23275
\n",
"
24278
\n",
"
22840
\n",
"
24129
\n",
"
24425
\n",
"
\n",
" \n",
"
\n",
"
4 rows × 116 columns
\n",
"
"
],
"text/plain": [
" cap-shape_b cap-shape_c cap-shape_f cap-shape_k cap-shape_s \\\n",
"count 25986 25986 25986 25986 25986 \n",
"unique 2 2 2 2 2 \n",
"top False False False False False \n",
"freq 23726 24509 19264 23074 24518 \n",
"\n",
" cap-shape_x cap-surface_f cap-surface_g cap-surface_s cap-surface_y \\\n",
"count 25986 25986 25986 25986 25986 \n",
"unique 2 2 2 2 2 \n",
"top False False False False False \n",
"freq 18312 19762 23825 19466 18384 \n",
"\n",
" ... population_s population_v population_y habitat_d habitat_g \\\n",
"count ... 25986 25986 25986 25986 25986 \n",
"unique ... 2 2 2 2 2 \n",
"top ... False False False False False \n",
"freq ... 22365 17577 21675 19413 21040 \n",
"\n",
" habitat_l habitat_m habitat_p habitat_u habitat_w \n",
"count 25986 25986 25986 25986 25986 \n",
"unique 2 2 2 2 2 \n",
"top False False False False False \n",
"freq 23275 24278 22840 24129 24425 \n",
"\n",
"[4 rows x 116 columns]"
]
},
"execution_count": 148,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import pandas as pd\n",
"from sklearn.model_selection import train_test_split\n",
"from sklearn.preprocessing import LabelEncoder\n",
"from sklearn.tree import DecisionTreeClassifier\n",
"from sklearn.metrics import accuracy_score, classification_report\n",
"\n",
"df = pd.read_csv(data_url)\n",
"\n",
"# drop the useless feature\n",
"df = df.drop('veil-type', axis=1)\n",
"\n",
"# drop the target from X -- and store it as y\n",
"X = df.drop('class', axis = 1)\n",
"y = df['class']\n",
"\n",
"# one-hot encode all columns at once\n",
"X = pd.get_dummies(X)\n",
"\n",
"# show it to me\n",
"X.describe()"
]
},
{
"cell_type": "markdown",
"id": "f2be1f0e",
"metadata": {},
"source": [
"### Decision Tree Classifier"
]
},
{
"cell_type": "markdown",
"id": "65e61b9e-d3b2-4aba-ac92-5d333bb5b802",
"metadata": {},
"source": [
"Our first model is a decision tree, which is one of the oldest algorithms for classifying observations. Before we create any models, we *always* create a train-test split so there is unseen testing data that wasn't available when the model was training."
]
},
{
"cell_type": "code",
"execution_count": 151,
"id": "1d56f782-edf6-4ea1-a615-c89f683ca68c",
"metadata": {
"scrolled": true
},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
0
\n",
"
1
\n",
"
2
\n",
"
3
\n",
"
4
\n",
"
5
\n",
"
6
\n",
"
7
\n",
"
8
\n",
"
9
\n",
"
...
\n",
"
5188
\n",
"
5189
\n",
"
5190
\n",
"
5191
\n",
"
5192
\n",
"
5193
\n",
"
5194
\n",
"
5195
\n",
"
5196
\n",
"
5197
\n",
"
\n",
" \n",
" \n",
"
\n",
"
0
\n",
"
p
\n",
"
p
\n",
"
e
\n",
"
p
\n",
"
p
\n",
"
p
\n",
"
e
\n",
"
e
\n",
"
e
\n",
"
p
\n",
"
...
\n",
"
p
\n",
"
e
\n",
"
p
\n",
"
e
\n",
"
p
\n",
"
p
\n",
"
p
\n",
"
p
\n",
"
e
\n",
"
p
\n",
"
\n",
"
\n",
"
1
\n",
"
p
\n",
"
e
\n",
"
e
\n",
"
p
\n",
"
p
\n",
"
p
\n",
"
e
\n",
"
p
\n",
"
e
\n",
"
p
\n",
"
...
\n",
"
p
\n",
"
p
\n",
"
p
\n",
"
e
\n",
"
e
\n",
"
e
\n",
"
p
\n",
"
p
\n",
"
e
\n",
"
p
\n",
"
\n",
" \n",
"
\n",
"
2 rows × 5198 columns
\n",
"
"
],
"text/plain": [
" 0 1 2 3 4 5 6 7 8 9 ... 5188 5189 5190 5191 \\\n",
"0 p p e p p p e e e p ... p e p e \n",
"1 p e e p p p e p e p ... p p p e \n",
"\n",
" 5192 5193 5194 5195 5196 5197 \n",
"0 p p p p e p \n",
"1 e e p p e p \n",
"\n",
"[2 rows x 5198 columns]"
]
},
"execution_count": 151,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Split the data into training and testing sets\n",
"X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n",
"\n",
"# Initialize the DecisionTreeClassifier\n",
"clf = DecisionTreeClassifier()\n",
"\n",
"# Fit the model\n",
"clf.fit(X_train, y_train)\n",
"\n",
"# Make predictions\n",
"y_pred = clf.predict(X_test)\n",
"results = pd.DataFrame([y_pred, y_test])\n",
"results"
]
},
{
"cell_type": "markdown",
"id": "e8a398ef-cfab-4c4a-9da8-c54e8007cb35",
"metadata": {},
"source": [
"There are many metrics for evaluating categorical models, and they are sometimes at odds. Accuracy is simplistic and obscures what could be more important -- are there more false positives or more false negatives? And which is more important? In a task to identify poisonous mushrooms, a false negative (labeling a 'p' as an 'e') is deadly. The \"recall\" on \"p\" below captures this value. This measures the percent of poisonous mushrooms you have correctly labeled as poisonous.\n",
"\n",
"Recall is not everything, though. You can easily get perfect \"poison\" recall by labeling every mushroom as poisonous! Precision measures the fraction of mushrooms you label as poisonous which actually are. \n",
"\n",
"The $F_1$-score is a type of geometric mean between precision and recall and strikes a bit of a balance between the two.\n",
"\n",
"In this notebook, look first at \"p-recall\", but keep an eye on the other metrics."
]
},
{
"cell_type": "code",
"execution_count": 153,
"id": "4126dbec-ab98-4a33-a19e-4cd5ad088cb6",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy: 0.6671796844940362\n",
"Classification Report:\n",
" precision recall f1-score support\n",
"\n",
" e 0.70 0.69 0.70 2873\n",
" p 0.62 0.64 0.63 2325\n",
"\n",
" accuracy 0.67 5198\n",
" macro avg 0.66 0.66 0.66 5198\n",
"weighted avg 0.67 0.67 0.67 5198\n",
"\n"
]
}
],
"source": [
"# Evaluate the model\n",
"accuracy = accuracy_score(y_test, y_pred)\n",
"print(f\"Accuracy: {accuracy}\")\n",
"\n",
"print(\"Classification Report:\")\n",
"print(classification_report(y_test, y_pred))"
]
},
{
"cell_type": "markdown",
"id": "f8466d20-62df-44d3-a115-d54d07e572d7",
"metadata": {},
"source": [
"A confusion matrix is a nice way to really show everything that a classifier is doing. The main diagonal are the numbers of correctly classified observations. The off-diagonals are errors. Unfortunately this simple version is unlabeled"
]
},
{
"cell_type": "code",
"execution_count": 157,
"id": "c641e9dc-e55a-4b53-8ee7-4b16f61ad6f7",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"array([[1978, 895],\n",
" [ 835, 1490]])"
]
},
"execution_count": 157,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n",
"\n",
"confusion_matrix(y_test, y_pred)"
]
},
{
"cell_type": "markdown",
"id": "d11a4123-27e4-4fe0-a349-11886cc5b7e8",
"metadata": {},
"source": [
"With a bit more work we can get a label."
]
},
{
"cell_type": "code",
"execution_count": 158,
"id": "3bc8d713-cb12-46c1-9ace-d977e48e5b68",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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",
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# Calculate the confusion matrix\n",
"cm = confusion_matrix(y_test, clf.predict(X_test))\n",
"cm_display = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=clf.classes_)\n",
"\n",
"# Plot the confusion matrix\n",
"fig, ax = plt.subplots(figsize=(8, 6))\n",
"sns.heatmap(cm, annot=True, fmt='d', cmap='Blues', xticklabels=clf.classes_, yticklabels=clf.classes_)\n",
"plt.xlabel('Predicted')\n",
"plt.ylabel('Actual')\n",
"plt.title('Confusion Matrix')\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"id": "390a17b1",
"metadata": {},
"source": [
"### Testing more methods"
]
},
{
"cell_type": "markdown",
"id": "209a30d2-00b0-4418-b188-fefad96c93cc",
"metadata": {},
"source": [
"First I'll define a helper method that takes any dataset and a classifier \"clf\" and\n",
"* fits the model to the training data\n",
"* applies the model to the test data\n",
"* gets an accuracy score and a classification report for the test data"
]
},
{
"cell_type": "code",
"execution_count": 159,
"id": "38933f3c",
"metadata": {},
"outputs": [],
"source": [
"def classifier_tryout(clf, X_train, y_train, X_test, y_test):\n",
"\tclf.fit(X_train, y_train)\n",
"\n",
"\t# Make predictions\n",
"\ty_pred = clf.predict(X_test)\n",
"\n",
"\t# Evaluate the model\n",
"\taccuracy = accuracy_score(y_test, y_pred)\n",
"\tprint(f\"Accuracy: {accuracy}\")\n",
"\n",
"\tprint(\"Classification Report:\")\n",
"\tprint(classification_report(y_test, y_pred))"
]
},
{
"cell_type": "markdown",
"id": "d5cdbfcd-beb8-4ace-86e2-b72ccb121a08",
"metadata": {},
"source": [
"In what follows, we run several very different models and compare their performance. We won't go into much detail about the models. But note how the scikit-learn API makes dealing with each of the models very similar"
]
},
{
"cell_type": "markdown",
"id": "06ef4cb1-23c9-468a-8803-0db9dc26ad0b",
"metadata": {},
"source": [
"### Random Forest"
]
},
{
"cell_type": "code",
"execution_count": 160,
"id": "876eb165-83e6-404d-9397-9ca9e0ed0985",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy: 0.7464409388226241\n",
"Classification Report:\n",
" precision recall f1-score support\n",
"\n",
" e 0.76 0.80 0.78 2873\n",
" p 0.73 0.69 0.71 2325\n",
"\n",
" accuracy 0.75 5198\n",
" macro avg 0.74 0.74 0.74 5198\n",
"weighted avg 0.75 0.75 0.75 5198\n",
"\n"
]
}
],
"source": [
"from sklearn.ensemble import RandomForestClassifier\n",
"\n",
"clf = RandomForestClassifier(random_state=42)\n",
"classifier_tryout(clf, X_train, y_train, X_test, y_test)"
]
},
{
"cell_type": "markdown",
"id": "1a69f38e-fbec-461b-9a67-6384638eb203",
"metadata": {},
"source": [
"### Support Vector Machines"
]
},
{
"cell_type": "code",
"execution_count": 116,
"id": "4decd3d2-2549-40f2-8d6f-6dd9a6e9ef6d",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy: 0.7504809542131589\n",
"Classification Report:\n",
" precision recall f1-score support\n",
"\n",
" e 0.76 0.79 0.78 2873\n",
" p 0.73 0.70 0.71 2325\n",
"\n",
" accuracy 0.75 5198\n",
" macro avg 0.75 0.75 0.75 5198\n",
"weighted avg 0.75 0.75 0.75 5198\n",
"\n"
]
}
],
"source": [
"from sklearn.svm import SVC\n",
"\n",
"# Initialize the RandomForestClassifier\n",
"clf = SVC(random_state=42, kernel='rbf')\n",
"classifier_tryout(clf, X_train, y_train, X_test, y_test)"
]
},
{
"cell_type": "markdown",
"id": "aed528e4-5585-4ec9-9fe6-399a1cb54256",
"metadata": {},
"source": [
"### Logistic Regression"
]
},
{
"cell_type": "code",
"execution_count": 101,
"id": "fd9ebd7b-48da-4652-a149-04bcf0bc015e",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy: 0.746056175452097\n",
"Classification Report:\n",
" precision recall f1-score support\n",
"\n",
" e 0.76 0.78 0.77 2873\n",
" p 0.72 0.70 0.71 2325\n",
"\n",
" accuracy 0.75 5198\n",
" macro avg 0.74 0.74 0.74 5198\n",
"weighted avg 0.75 0.75 0.75 5198\n",
"\n"
]
}
],
"source": [
"from sklearn.linear_model import LogisticRegression\n",
"\n",
"# Initialize the LogisticRegression\n",
"clf = LogisticRegression(random_state=42)\n",
"classifier_tryout(clf, X_train, y_train, X_test, y_test)"
]
},
{
"cell_type": "markdown",
"id": "eb802b79-197a-411d-aec9-c203c956ef30",
"metadata": {},
"source": [
"### k-Nearest Neighbors"
]
},
{
"cell_type": "code",
"execution_count": 102,
"id": "31601aa2-588b-4997-aee2-dbc36e9ea45e",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy: 0.6854559445940747\n",
"Classification Report:\n",
" precision recall f1-score support\n",
"\n",
" e 0.70 0.76 0.73 2873\n",
" p 0.67 0.60 0.63 2325\n",
"\n",
" accuracy 0.69 5198\n",
" macro avg 0.68 0.68 0.68 5198\n",
"weighted avg 0.68 0.69 0.68 5198\n",
"\n"
]
}
],
"source": [
"from sklearn.neighbors import KNeighborsClassifier\n",
"\n",
"# Initialize the KNeighborsClassifier\n",
"clf = KNeighborsClassifier(weights='uniform')\n",
"classifier_tryout(clf, X_train, y_train, X_test, y_test)"
]
},
{
"cell_type": "markdown",
"id": "f9bf489b-b3f8-45c3-8d4d-979e76eb1f7c",
"metadata": {},
"source": [
"### GradientBoost "
]
},
{
"cell_type": "code",
"execution_count": 103,
"id": "3f393c0c-534d-44d7-8b12-3f5659a5117b",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy: 0.7518276260100039\n",
"Classification Report:\n",
" precision recall f1-score support\n",
"\n",
" e 0.77 0.79 0.78 2873\n",
" p 0.73 0.70 0.72 2325\n",
"\n",
" accuracy 0.75 5198\n",
" macro avg 0.75 0.75 0.75 5198\n",
"weighted avg 0.75 0.75 0.75 5198\n",
"\n"
]
}
],
"source": [
"from sklearn.ensemble import GradientBoostingClassifier\n",
"\n",
"# Initialize the GradientBoostingClassifier\n",
"clf = GradientBoostingClassifier(random_state=42)\n",
"classifier_tryout(clf, X_train, y_train, X_test, y_test)"
]
},
{
"cell_type": "markdown",
"id": "74f7245a-3478-4c53-afbf-53c2a9b3e0f4",
"metadata": {},
"source": [
"### Neural Network"
]
},
{
"cell_type": "code",
"execution_count": 104,
"id": "786d1642-2aec-4987-bad9-afe56337cae2",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy: 0.7052712581762216\n",
"Classification Report:\n",
" precision recall f1-score support\n",
"\n",
" e 0.73 0.73 0.73 2873\n",
" p 0.67 0.67 0.67 2325\n",
"\n",
" accuracy 0.71 5198\n",
" macro avg 0.70 0.70 0.70 5198\n",
"weighted avg 0.71 0.71 0.71 5198\n",
"\n"
]
}
],
"source": [
"from sklearn.neural_network import MLPClassifier\n",
"\n",
"# Initialize the MLPClassifier\n",
"clf = MLPClassifier(random_state=42, hidden_layer_sizes=(1000,10,), learning_rate='adaptive')\n",
"classifier_tryout(clf, X_train, y_train, X_test, y_test)"
]
},
{
"cell_type": "markdown",
"id": "366f9ce3-2956-4dfd-a960-92697dd3e669",
"metadata": {},
"source": [
"## Conclusion"
]
},
{
"cell_type": "markdown",
"id": "e1c5e40a-a3ca-4a21-b9db-a3f6394d49a6",
"metadata": {},
"source": [
"This notebook gave a quick overview of the full data analysis workflow on a real world dataset. Data ingesting and cleaning; exploratory data analysis; modeling. There is much we can do next by trying to optimize models, but that is beyond the scope of this notebook. Also, we only looked at categorical data here -- continuous/quantitative data is a bit different but not much (linear regression is the first stop with continuous data.)\n",
"\n",
"Which model above do you think is the best? Which one would you start with to try to get even better results?\n",
"\n",
"As an application, find your own dataset and try to mirror the process we took here. Not all the steps in this notebook will apply directly to whatever dataset you find, but see what you can come up with and how well you can model the target in your data!"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.11.9"
}
},
"nbformat": 4,
"nbformat_minor": 5
}