return

0 - Classification pour "Place de marché"

Classification de produits par images et mots

Consignes:

Vous analyserez le jeu de données en réalisant un prétraitement des images et des descriptions des produits, une réduction de dimension, puis un clustering. Les résultats du clustering seront présentés sous la forme d’une représentation en deux dimensions à déterminer, qui ’illustrera le fait que les caractéristiques extraites permettent de regrouper des produits de même catégorie.

La représentation graphique vous aidera à convaincre Linda que cette approche de modélisation permettra bien de regrouper des produits de même catégorie.

---

1 - Librairies et fichiers de données

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

#For ImageClassification
from skimage.io import imread
from IPython.display import Image
import os
import cv2
from sklearn.cluster import MiniBatchKMeans
from xgboost import XGBRFClassifier
from sklearn.metrics import accuracy_score

#for WordClassification
import re
import nltk
from nltk.tokenize import word_tokenize as wt 
from nltk.corpus import stopwords
from nltk.stem.porter import PorterStemmer
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.naive_bayes import GaussianNB
from sklearn.metrics import confusion_matrix, classification_report, accuracy_score

import warnings
warnings.simplefilter(action='ignore', category=FutureWarning)

data = pd.read_csv('flipkart_com-ecommerce_sample_1050.csv')
print(data.shape)
display(data.head(2))

(1050, 15)

	uniq_id	crawl_timestamp	product_url	product_name	product_category_tree	pid	retail_price	discounted_price	image	is_FK_Advantage_product	description	product_rating	overall_rating	brand	product_specifications
0	55b85ea15a1536d46b7190ad6fff8ce7	2016-04-30 03:22:56 +0000	http://www.flipkart.com/elegance-polyester-mul...	Elegance Polyester Multicolor Abstract Eyelet ...	["Home Furnishing >> Curtains & Accessories >>...	CRNEG7BKMFFYHQ8Z	1899.0	899.0	55b85ea15a1536d46b7190ad6fff8ce7.jpg	False	Key Features of Elegance Polyester Multicolor ...	No rating available	No rating available	Elegance	{"product_specification"=>[{"key"=>"Brand", "v...
1	7b72c92c2f6c40268628ec5f14c6d590	2016-04-30 03:22:56 +0000	http://www.flipkart.com/sathiyas-cotton-bath-t...	Sathiyas Cotton Bath Towel	["Baby Care >> Baby Bath & Skin >> Baby Bath T...	BTWEGFZHGBXPHZUH	600.0	449.0	7b72c92c2f6c40268628ec5f14c6d590.jpg	False	Specifications of Sathiyas Cotton Bath Towel (...	No rating available	No rating available	Sathiyas	{"product_specification"=>[{"key"=>"Machine Wa...

display(data.info())
display(data.describe(include='all'))

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1050 entries, 0 to 1049
Data columns (total 15 columns):
 #   Column                   Non-Null Count  Dtype  
---  ------                   --------------  -----  
 0   uniq_id                  1050 non-null   object 
 1   crawl_timestamp          1050 non-null   object 
 2   product_url              1050 non-null   object 
 3   product_name             1050 non-null   object 
 4   product_category_tree    1050 non-null   object 
 5   pid                      1050 non-null   object 
 6   retail_price             1049 non-null   float64
 7   discounted_price         1049 non-null   float64
 8   image                    1050 non-null   object 
 9   is_FK_Advantage_product  1050 non-null   bool   
 10  description              1050 non-null   object 
 11  product_rating           1050 non-null   object 
 12  overall_rating           1050 non-null   object 
 13  brand                    712 non-null    object 
 14  product_specifications   1049 non-null   object 
dtypes: bool(1), float64(2), object(12)
memory usage: 116.0+ KB

None

	uniq_id	crawl_timestamp	product_url	product_name	product_category_tree	pid	retail_price	discounted_price	image	is_FK_Advantage_product	description	product_rating	overall_rating	brand	product_specifications
count	1050	1050	1050	1050	1050	1050	1049.000000	1049.000000	1050	1050	1050	1050	1050	712	1049
unique	1050	149	1050	1050	642	1050	NaN	NaN	1050	2	1050	27	27	490	984
top	aa82b75da7579007963e53b6f818281b	2015-12-01 12:40:44 +0000	http://www.flipkart.com/wallmantra-christ-rede...	Vitamins Solid Baby Girl's Basic Shorts	["Home Furnishing >> Bed Linen >> Blankets, Qu...	NPYEYC9ZV2CUG9H3	NaN	NaN	d136aa676ef52b09eab65762940957fe.jpg	False	arnavs Multi1 Bottle Opener Set Price: Rs. 250...	No rating available	No rating available	PRINT SHAPES	{"product_specification"=>[{"key"=>"Type", "va...
freq	1	150	1	1	56	1	NaN	NaN	1	993	1	889	889	11	22
mean	NaN	NaN	NaN	NaN	NaN	NaN	2186.197331	1584.527169	NaN	NaN	NaN	NaN	NaN	NaN	NaN
std	NaN	NaN	NaN	NaN	NaN	NaN	7639.229411	7475.099680	NaN	NaN	NaN	NaN	NaN	NaN	NaN
min	NaN	NaN	NaN	NaN	NaN	NaN	35.000000	35.000000	NaN	NaN	NaN	NaN	NaN	NaN	NaN
25%	NaN	NaN	NaN	NaN	NaN	NaN	555.000000	340.000000	NaN	NaN	NaN	NaN	NaN	NaN	NaN
50%	NaN	NaN	NaN	NaN	NaN	NaN	999.000000	600.000000	NaN	NaN	NaN	NaN	NaN	NaN	NaN
75%	NaN	NaN	NaN	NaN	NaN	NaN	1999.000000	1199.000000	NaN	NaN	NaN	NaN	NaN	NaN	NaN
max	NaN	NaN	NaN	NaN	NaN	NaN	201000.000000	201000.000000	NaN	NaN	NaN	NaN	NaN	NaN	NaN

#valeur manquante
data.isna().sum()

uniq_id                      0
crawl_timestamp              0
product_url                  0
product_name                 0
product_category_tree        0
pid                          0
retail_price                 1
discounted_price             1
image                        0
is_FK_Advantage_product      0
description                  0
product_rating               0
overall_rating               0
brand                      338
product_specifications       1
dtype: int64

data['retail_price'].fillna(data['retail_price'].mean(),inplace=True)
data['discounted_price'].fillna(data['discounted_price'].mean(),inplace=True)
data['brand'].fillna("",inplace=True)
data['product_specifications'].fillna("",inplace=True)
#valeur manquante
data.isna().sum()

uniq_id                    0
crawl_timestamp            0
product_url                0
product_name               0
product_category_tree      0
pid                        0
retail_price               0
discounted_price           0
image                      0
is_FK_Advantage_product    0
description                0
product_rating             0
overall_rating             0
brand                      0
product_specifications     0
dtype: int64

# Exemple de la catégorie du produit 0
print(data["product_category_tree"][0])

["Home Furnishing >> Curtains & Accessories >> Curtains >> Elegance Polyester Multicolor Abstract Eyelet Do..."]

# On récupère la colomne "product_category_tree"
list_categories_1=[]
list_categories_2 =[]
for txt in data["product_category_tree"] :
    list_categories_1.append(txt.split(">>")[0].split("\"")[1].strip()) 
    list_categories_2.append(txt.split(">>")[1].strip())   
    
    
data["categories_niv1"] = pd.Series(list_categories_1)

data["categories_niv2"] = pd.Series(list_categories_2)

print(data["categories_niv1"].nunique()," catégories de 1er niveau")

print('----')
for i in data["categories_niv1"].unique():
    print(i,len(data[data["categories_niv1"]==i]))
print("-> en moyenne nous avons ~150 images par catégorie au niveau  de profondeur 1")
    

# Barplot
fig = plt.figure(1, figsize=(15, 5))
plt.subplot(1, 2, 1)
plt.title("Nombre de produits FlipKart dans chaque catégorie au niveau 1")

sns.barplot(data["categories_niv1"].value_counts().index, data["categories_niv1"].value_counts().values, orient="v")
plt.xticks(rotation=45)
plt.show()

7  catégories de 1er niveau
----
Home Furnishing 150
Baby Care 150
Watches 150
Home Decor & Festive Needs 150
Kitchen & Dining 150
Beauty and Personal Care 150
Computers 150
-> en moyenne nous avons ~150 images par catégorie au niveau  de profondeur 1

print(data["categories_niv2"].nunique()," categories au 2ième niveau de catégories")
print('----')

print("-> Le nombre d'image par catégorie au niveau de profondeur 2 diffère énormément")
fig = plt.figure(1, figsize=(45, 10))
# Barplot
plt.subplot(1, 2, 1)
plt.title("Nombre de produits dans chaque catégorie au niveau 2")

sns.barplot(data["categories_niv2"].value_counts().index, data["categories_niv2"].value_counts().values, orient="v")
plt.xticks(rotation=60)
plt.show()

63  categories au 2ième niveau de catégories
----
-> Le nombre d'image par catégorie au niveau de profondeur 2 diffère énormément

⤷ On préferera les catégorie de niveau 1 pour le travail de classification

---

df=data[['categories_niv1','image']].copy()
dataWCL = data[['categories_niv1','description']].copy()

#export du ficher de data préparé.
data[['categories_niv1','image','description']].to_csv('data_Img_cat_desc.csv',index=False)

2 - Classification sur image avec algorithme ORB

Bag-of- features model with ORB descriptors

df.drop(343,axis=0,inplace=True)

Extraction des keypoints de chaque images

dico=[] #contients toute les images sous formes de tableau "applatit"
target_arr=[] #pour chaque image sa catégorie sous forme de nombre(baby care=0,wactches=6,...)
#Peux prendre vraiment bcp de temps(jusqu'à 18min).... mais ça marche
datadir='./Images/' 
#nl,to delete:
list_image=[]

sift = cv2.xfeatures2d.SIFT_create()
orb = cv2.ORB_create()
for img in os.listdir(datadir):
    list_image.append(img)
    img_array=imread(os.path.join(datadir,img))
    #img_resized=resize(img_array,(150,150,3))
    gray_img = cv2.cvtColor(img_array, cv2.COLOR_BGR2GRAY)
    kp, des = orb.detectAndCompute(gray_img, None)
    for d in des:
        dico.append(d)

C:\Users\malik\anaconda3\lib\site-packages\PIL\Image.py:2834: DecompressionBombWarning: Image size (93680328 pixels) exceeds limit of 89478485 pixels, could be decompression bomb DOS attack.
  warnings.warn(

Les catégories

Transformations de catégories types catégorielles en type numériques

df.categories_niv1.unique()

array(['Home Furnishing', 'Baby Care', 'Watches',
       'Home Decor & Festive Needs', 'Kitchen & Dining',
       'Beauty and Personal Care', 'Computers'], dtype=object)

df= df[df['image'].isin(list_image)].copy()

listOfCategories = ['Baby Care','Beauty and Personal Care','Computers','Home Decor & Festive Needs','Home Furnishing','Kitchen & Dining','Watches']

df.replace(['Baby Care','Beauty and Personal Care','Computers','Home Decor & Festive Needs','Home Furnishing','Kitchen & Dining','Watches'],
           [0,1,2,3,4,5,6],inplace=True)
#2 photos n'ont pas de categories, on regardes les photos et on décide de mettre la catégories 5 
df.fillna(value=5,inplace=True)
df.categorie=df.categories_niv1.astype(int).copy()

<ipython-input-14-998fcfad0181>:9: UserWarning: Pandas doesn't allow columns to be created via a new attribute name - see https://pandas.pydata.org/pandas-docs/stable/indexing.html#attribute-access
  df.categorie=df.categories_niv1.astype(int).copy()

Clustering

Nous avons à présent un grand nombre de descripteurs par images dans un tableau. Mais il est impossible de tous les utiliser. Il faut donc regrouper en cluster.

k = np.size(listOfCategories) * 10
#nl,nb image*3
batch_size = np.size(os.listdir(datadir)) * 3
kmeans = MiniBatchKMeans(n_clusters=k, batch_size=batch_size, verbose=1).fit(dico)

Init 1/3 with method: k-means++
Inertia for init 1/3: 1044834789.251844
Init 2/3 with method: k-means++
Inertia for init 2/3: 1043343231.234659
Init 3/3 with method: k-means++
Inertia for init 3/3: 1042087132.457279
Minibatch iteration 1/16100: mean batch inertia: 110913.367349, ewa inertia: 110913.367349 
Minibatch iteration 2/16100: mean batch inertia: 110372.006269, ewa inertia: 110906.609489 
Minibatch iteration 3/16100: mean batch inertia: 110153.263719, ewa inertia: 110897.205404 
Minibatch iteration 4/16100: mean batch inertia: 110683.445215, ewa inertia: 110894.537016 
Minibatch iteration 5/16100: mean batch inertia: 109568.849088, ewa inertia: 110877.988332 
Minibatch iteration 6/16100: mean batch inertia: 109776.100173, ewa inertia: 110864.233360 
Minibatch iteration 7/16100: mean batch inertia: 109093.429560, ewa inertia: 110842.128254 
Minibatch iteration 8/16100: mean batch inertia: 107733.716531, ewa inertia: 110803.325664 
Minibatch iteration 9/16100: mean batch inertia: 108747.181610, ewa inertia: 110777.658628 
Minibatch iteration 10/16100: mean batch inertia: 109152.138814, ewa inertia: 110757.367115 
Minibatch iteration 11/16100: mean batch inertia: 108529.178729, ewa inertia: 110729.552434 
Minibatch iteration 12/16100: mean batch inertia: 108179.662171, ewa inertia: 110697.721919 
Minibatch iteration 13/16100: mean batch inertia: 108181.495067, ewa inertia: 110666.311627 
Minibatch iteration 14/16100: mean batch inertia: 108018.247772, ewa inertia: 110633.255602 
Minibatch iteration 15/16100: mean batch inertia: 108458.568501, ewa inertia: 110606.108783 
Minibatch iteration 16/16100: mean batch inertia: 108360.145374, ewa inertia: 110578.072215 
Minibatch iteration 17/16100: mean batch inertia: 108089.964859, ewa inertia: 110547.012942 
Minibatch iteration 18/16100: mean batch inertia: 108382.660548, ewa inertia: 110519.995131 
Minibatch iteration 19/16100: mean batch inertia: 108256.472462, ewa inertia: 110491.739369 
Minibatch iteration 20/16100: mean batch inertia: 106666.993356, ewa inertia: 110443.994713 
Minibatch iteration 21/16100: mean batch inertia: 107567.317727, ewa inertia: 110408.084889 
Minibatch iteration 22/16100: mean batch inertia: 108638.389276, ewa inertia: 110385.993616 
Minibatch iteration 23/16100: mean batch inertia: 107909.897174, ewa inertia: 110355.084276 
Minibatch iteration 24/16100: mean batch inertia: 107650.559392, ewa inertia: 110321.323443 
Minibatch iteration 25/16100: mean batch inertia: 108247.026777, ewa inertia: 110295.429806 
Minibatch iteration 26/16100: mean batch inertia: 107608.693223, ewa inertia: 110261.891026 
Minibatch iteration 27/16100: mean batch inertia: 108020.993345, ewa inertia: 110233.917694 
Minibatch iteration 28/16100: mean batch inertia: 107895.646042, ewa inertia: 110204.728834 
Minibatch iteration 29/16100: mean batch inertia: 108722.550297, ewa inertia: 110186.226662 
Minibatch iteration 30/16100: mean batch inertia: 106944.883666, ewa inertia: 110145.764679 
Minibatch iteration 31/16100: mean batch inertia: 107572.725051, ewa inertia: 110113.645189 
Minibatch iteration 32/16100: mean batch inertia: 107173.722245, ewa inertia: 110076.945860 
Minibatch iteration 33/16100: mean batch inertia: 107874.586842, ewa inertia: 110049.453609 
Minibatch iteration 34/16100: mean batch inertia: 107575.698422, ewa inertia: 110018.573495 
Minibatch iteration 35/16100: mean batch inertia: 107454.186594, ewa inertia: 109986.562018 
Minibatch iteration 36/16100: mean batch inertia: 106603.685487, ewa inertia: 109944.333258 
Minibatch iteration 37/16100: mean batch inertia: 108111.589497, ewa inertia: 109921.454949 
Minibatch iteration 38/16100: mean batch inertia: 106711.609243, ewa inertia: 109881.386150 
Minibatch iteration 39/16100: mean batch inertia: 107240.926323, ewa inertia: 109848.425046 
Minibatch iteration 40/16100: mean batch inertia: 107084.411573, ewa inertia: 109813.921612 
Minibatch iteration 41/16100: mean batch inertia: 106596.858463, ewa inertia: 109773.762716 
Minibatch iteration 42/16100: mean batch inertia: 108070.288125, ewa inertia: 109752.498086 
Minibatch iteration 43/16100: mean batch inertia: 108005.148357, ewa inertia: 109730.685758 
Minibatch iteration 44/16100: mean batch inertia: 107908.390393, ewa inertia: 109707.937877 
Minibatch iteration 45/16100: mean batch inertia: 106649.107947, ewa inertia: 109669.754222 
Minibatch iteration 46/16100: mean batch inertia: 107146.966465, ewa inertia: 109638.262030 
Minibatch iteration 47/16100: mean batch inertia: 107088.943773, ewa inertia: 109606.438655 
Minibatch iteration 48/16100: mean batch inertia: 107535.813183, ewa inertia: 109580.590847 
Minibatch iteration 49/16100: mean batch inertia: 107034.217562, ewa inertia: 109548.804235 
Minibatch iteration 50/16100: mean batch inertia: 107858.848478, ewa inertia: 109527.708361 
Minibatch iteration 51/16100: mean batch inertia: 107416.426550, ewa inertia: 109501.353036 
Minibatch iteration 52/16100: mean batch inertia: 107754.780427, ewa inertia: 109479.550409 
Minibatch iteration 53/16100: mean batch inertia: 107118.049225, ewa inertia: 109450.071572 
Minibatch iteration 54/16100: mean batch inertia: 107086.307857, ewa inertia: 109420.564492 
Minibatch iteration 55/16100: mean batch inertia: 107496.168547, ewa inertia: 109396.542081 
Minibatch iteration 56/16100: mean batch inertia: 106586.888061, ewa inertia: 109361.468911 
Minibatch iteration 57/16100: mean batch inertia: 107633.376120, ewa inertia: 109339.896969 
Minibatch iteration 58/16100: mean batch inertia: 106882.382340, ewa inertia: 109309.219588 
Minibatch iteration 59/16100: mean batch inertia: 107096.361542, ewa inertia: 109281.596277 
Minibatch iteration 60/16100: mean batch inertia: 106358.482530, ewa inertia: 109245.106779 
Minibatch iteration 61/16100: mean batch inertia: 106471.240056, ewa inertia: 109210.480345 
Minibatch iteration 62/16100: mean batch inertia: 107291.966093, ewa inertia: 109186.531355 
Minibatch iteration 63/16100: mean batch inertia: 107293.851525, ewa inertia: 109162.904858 
Minibatch iteration 64/16100: mean batch inertia: 107180.043426, ewa inertia: 109138.152617 
Minibatch iteration 65/16100: mean batch inertia: 107372.949396, ewa inertia: 109116.117422 
Minibatch iteration 66/16100: mean batch inertia: 106576.840114, ewa inertia: 109084.419390 
Minibatch iteration 67/16100: mean batch inertia: 107740.850236, ewa inertia: 109067.647492 
Minibatch iteration 68/16100: mean batch inertia: 107716.487321, ewa inertia: 109050.780835 
Minibatch iteration 69/16100: mean batch inertia: 107059.715263, ewa inertia: 109025.926180 
Minibatch iteration 70/16100: mean batch inertia: 106970.967912, ewa inertia: 109000.273947 
Minibatch iteration 71/16100: mean batch inertia: 107056.335132, ewa inertia: 108976.007580 
Minibatch iteration 72/16100: mean batch inertia: 106465.605654, ewa inertia: 108944.670001 
Minibatch iteration 73/16100: mean batch inertia: 107586.670392, ewa inertia: 108927.717967 
Minibatch iteration 74/16100: mean batch inertia: 106027.330122, ewa inertia: 108891.512159 
Minibatch iteration 75/16100: mean batch inertia: 107314.938536, ewa inertia: 108871.831645 
Minibatch iteration 76/16100: mean batch inertia: 106919.763868, ewa inertia: 108847.463803 
Minibatch iteration 77/16100: mean batch inertia: 106517.885964, ewa inertia: 108818.383468 
Minibatch iteration 78/16100: mean batch inertia: 106536.452710, ewa inertia: 108789.897917 
Minibatch iteration 79/16100: mean batch inertia: 107987.305216, ewa inertia: 108779.879078 
Minibatch iteration 80/16100: mean batch inertia: 107378.832307, ewa inertia: 108762.389682 
Minibatch iteration 81/16100: mean batch inertia: 106303.037959, ewa inertia: 108731.689368 
Minibatch iteration 82/16100: mean batch inertia: 107414.694707, ewa inertia: 108715.249202 
Minibatch iteration 83/16100: mean batch inertia: 106525.564608, ewa inertia: 108687.915168 
Minibatch iteration 84/16100: mean batch inertia: 106312.743413, ewa inertia: 108658.265680 
Minibatch iteration 85/16100: mean batch inertia: 107716.727694, ewa inertia: 108646.512375 
Minibatch iteration 86/16100: mean batch inertia: 106282.518878, ewa inertia: 108617.002426 
Minibatch iteration 87/16100: mean batch inertia: 107267.019909, ewa inertia: 108600.150470 
Minibatch iteration 88/16100: mean batch inertia: 107812.290732, ewa inertia: 108590.315545 
Minibatch iteration 89/16100: mean batch inertia: 107699.622176, ewa inertia: 108579.196937 
Minibatch iteration 90/16100: mean batch inertia: 107193.137811, ewa inertia: 108561.894633 
Minibatch iteration 91/16100: mean batch inertia: 106094.421599, ewa inertia: 108531.092940 
Minibatch iteration 92/16100: mean batch inertia: 107622.305631, ewa inertia: 108519.748465 
Minibatch iteration 93/16100: mean batch inertia: 106633.071337, ewa inertia: 108496.196900 
Minibatch iteration 94/16100: mean batch inertia: 106176.731706, ewa inertia: 108467.242803 
Minibatch iteration 95/16100: mean batch inertia: 107210.676714, ewa inertia: 108451.556973 
Minibatch iteration 96/16100: mean batch inertia: 107229.166561, ewa inertia: 108436.297760 
Minibatch iteration 97/16100: mean batch inertia: 107110.109364, ewa inertia: 108419.742829 
Minibatch iteration 98/16100: mean batch inertia: 107004.779689, ewa inertia: 108402.079713 
Minibatch iteration 99/16100: mean batch inertia: 106561.480961, ewa inertia: 108379.103350 
Minibatch iteration 100/16100: mean batch inertia: 107288.791623, ewa inertia: 108365.492888 
Minibatch iteration 101/16100: mean batch inertia: 107720.628737, ewa inertia: 108357.442989 
Minibatch iteration 102/16100: mean batch inertia: 107048.903303, ewa inertia: 108341.108368 
Minibatch iteration 103/16100: mean batch inertia: 107794.422316, ewa inertia: 108334.284036 
Minibatch iteration 104/16100: mean batch inertia: 107160.777291, ewa inertia: 108319.635043 
Minibatch iteration 105/16100: mean batch inertia: 106011.084926, ewa inertia: 108290.817199 
Minibatch iteration 106/16100: mean batch inertia: 107893.414567, ewa inertia: 108285.856385 
Minibatch iteration 107/16100: mean batch inertia: 106503.553988, ewa inertia: 108263.607741 
Minibatch iteration 108/16100: mean batch inertia: 106636.356337, ewa inertia: 108243.294611 
Minibatch iteration 109/16100: mean batch inertia: 106772.232975, ewa inertia: 108224.931213 
Minibatch iteration 110/16100: mean batch inertia: 106573.224227, ewa inertia: 108204.312803 
Minibatch iteration 111/16100: mean batch inertia: 106987.097650, ewa inertia: 108189.118194 
Minibatch iteration 112/16100: mean batch inertia: 107530.982951, ewa inertia: 108180.902631 
Minibatch iteration 113/16100: mean batch inertia: 107412.703887, ewa inertia: 108171.313136 
Minibatch iteration 114/16100: mean batch inertia: 107089.029439, ewa inertia: 108157.802889 
Minibatch iteration 115/16100: mean batch inertia: 106874.001848, ewa inertia: 108141.777082 
Minibatch iteration 116/16100: mean batch inertia: 106935.484007, ewa inertia: 108126.718814 
Minibatch iteration 117/16100: mean batch inertia: 105372.985593, ewa inertia: 108092.343709 
Minibatch iteration 118/16100: mean batch inertia: 106896.148037, ewa inertia: 108077.411488 
Minibatch iteration 119/16100: mean batch inertia: 107443.633554, ewa inertia: 108069.499980 
Minibatch iteration 120/16100: mean batch inertia: 107835.087329, ewa inertia: 108066.573785 
Minibatch iteration 121/16100: mean batch inertia: 107375.857184, ewa inertia: 108057.951506 
Minibatch iteration 122/16100: mean batch inertia: 106869.980358, ewa inertia: 108043.121953 
Minibatch iteration 123/16100: mean batch inertia: 106874.930151, ewa inertia: 108028.539307 
Minibatch iteration 124/16100: mean batch inertia: 106621.721305, ewa inertia: 108010.977869 
Minibatch iteration 125/16100: mean batch inertia: 106269.244434, ewa inertia: 107989.235650 
Minibatch iteration 126/16100: mean batch inertia: 107242.837527, ewa inertia: 107979.918293 
Minibatch iteration 127/16100: mean batch inertia: 106237.190783, ewa inertia: 107958.163665 
Minibatch iteration 128/16100: mean batch inertia: 106795.365715, ewa inertia: 107943.648351 
Minibatch iteration 129/16100: mean batch inertia: 106774.445806, ewa inertia: 107929.053088 
Minibatch iteration 130/16100: mean batch inertia: 106559.248513, ewa inertia: 107911.953692 
Minibatch iteration 131/16100: mean batch inertia: 106262.492252, ewa inertia: 107891.363312 
Minibatch iteration 132/16100: mean batch inertia: 106880.531159, ewa inertia: 107878.745002 
Minibatch iteration 133/16100: mean batch inertia: 107015.543422, ewa inertia: 107867.969577 
Minibatch iteration 134/16100: mean batch inertia: 107912.830892, ewa inertia: 107868.529585 
Minibatch iteration 135/16100: mean batch inertia: 106584.283962, ewa inertia: 107852.498228 
Minibatch iteration 136/16100: mean batch inertia: 107082.016089, ewa inertia: 107842.880229 
Minibatch iteration 137/16100: mean batch inertia: 107119.602122, ewa inertia: 107833.851482 
Minibatch iteration 138/16100: mean batch inertia: 107424.251517, ewa inertia: 107828.738408 
Minibatch iteration 139/16100: mean batch inertia: 106592.655027, ewa inertia: 107813.308265 
Minibatch iteration 140/16100: mean batch inertia: 106223.747526, ewa inertia: 107793.465632 
Minibatch iteration 141/16100: mean batch inertia: 107395.277417, ewa inertia: 107788.495012 
Minibatch iteration 142/16100: mean batch inertia: 106958.042099, ewa inertia: 107778.128391 
Minibatch iteration 143/16100: mean batch inertia: 107972.114371, ewa inertia: 107780.549936 
Minibatch iteration 144/16100: mean batch inertia: 107670.762523, ewa inertia: 107779.179450 
Minibatch iteration 145/16100: mean batch inertia: 106968.080786, ewa inertia: 107769.054431 
Minibatch iteration 146/16100: mean batch inertia: 106727.216383, ewa inertia: 107756.049070 
Minibatch iteration 147/16100: mean batch inertia: 107476.620248, ewa inertia: 107752.560935 
Minibatch iteration 148/16100: mean batch inertia: 107164.210470, ewa inertia: 107745.216502 
Minibatch iteration 149/16100: mean batch inertia: 108709.394601, ewa inertia: 107757.252426 
Minibatch iteration 150/16100: mean batch inertia: 106461.417850, ewa inertia: 107741.076403 
Minibatch iteration 151/16100: mean batch inertia: 106461.899882, ewa inertia: 107725.108325 
Minibatch iteration 152/16100: mean batch inertia: 107549.569862, ewa inertia: 107722.917062 
Minibatch iteration 153/16100: mean batch inertia: 106737.030610, ewa inertia: 107710.610151 
Minibatch iteration 154/16100: mean batch inertia: 107322.902658, ewa inertia: 107705.770362 
Minibatch iteration 155/16100: mean batch inertia: 106231.397881, ewa inertia: 107687.365635 
Minibatch iteration 156/16100: mean batch inertia: 106945.169337, ewa inertia: 107678.100730 
Minibatch iteration 157/16100: mean batch inertia: 107751.105037, ewa inertia: 107679.012050 
Minibatch iteration 158/16100: mean batch inertia: 107516.881094, ewa inertia: 107676.988154 
Minibatch iteration 159/16100: mean batch inertia: 107369.247210, ewa inertia: 107673.146595 
Minibatch iteration 160/16100: mean batch inertia: 106760.105494, ewa inertia: 107661.749019 
Minibatch iteration 161/16100: mean batch inertia: 106881.885883, ewa inertia: 107652.013916 
Minibatch iteration 162/16100: mean batch inertia: 107470.146283, ewa inertia: 107649.743646 
Minibatch iteration 163/16100: mean batch inertia: 107389.249778, ewa inertia: 107646.491877 
Minibatch iteration 164/16100: mean batch inertia: 107188.157369, ewa inertia: 107640.770445 
Minibatch iteration 165/16100: mean batch inertia: 107141.706235, ewa inertia: 107634.540580 
Minibatch iteration 166/16100: mean batch inertia: 107315.357853, ewa inertia: 107630.556193 
Minibatch iteration 167/16100: mean batch inertia: 107024.314650, ewa inertia: 107622.988424 
Minibatch iteration 168/16100: mean batch inertia: 107400.506398, ewa inertia: 107620.211160 
Minibatch iteration 169/16100: mean batch inertia: 106460.860863, ewa inertia: 107605.738884 
Minibatch iteration 170/16100: mean batch inertia: 107189.344950, ewa inertia: 107600.541000 
Minibatch iteration 171/16100: mean batch inertia: 107263.055198, ewa inertia: 107596.328134 
Minibatch iteration 172/16100: mean batch inertia: 107028.450901, ewa inertia: 107589.239270 
Minibatch iteration 173/16100: mean batch inertia: 107032.240633, ewa inertia: 107582.286204 
Minibatch iteration 174/16100: mean batch inertia: 107223.186349, ewa inertia: 107577.803528 
Minibatch iteration 175/16100: mean batch inertia: 106710.489366, ewa inertia: 107566.976765 
Minibatch iteration 176/16100: mean batch inertia: 106829.449335, ewa inertia: 107557.770142 
Minibatch iteration 177/16100: mean batch inertia: 106717.267867, ewa inertia: 107547.278075 
Minibatch iteration 178/16100: mean batch inertia: 107499.603046, ewa inertia: 107546.682943 
Minibatch iteration 179/16100: mean batch inertia: 106851.379148, ewa inertia: 107538.003402 
Minibatch iteration 180/16100: mean batch inertia: 106818.073433, ewa inertia: 107529.016450 
Minibatch iteration 181/16100: mean batch inertia: 106520.289412, ewa inertia: 107516.424417 
Minibatch iteration 182/16100: mean batch inertia: 106754.871558, ewa inertia: 107506.917883 
Minibatch iteration 183/16100: mean batch inertia: 107144.610351, ewa inertia: 107502.395165 
Minibatch iteration 184/16100: mean batch inertia: 106512.632551, ewa inertia: 107490.039867 
Minibatch iteration 185/16100: mean batch inertia: 107921.976262, ewa inertia: 107495.431769 
Minibatch iteration 186/16100: mean batch inertia: 106756.404883, ewa inertia: 107486.206428 
Minibatch iteration 187/16100: mean batch inertia: 106504.230971, ewa inertia: 107473.948338 
Minibatch iteration 188/16100: mean batch inertia: 106701.042609, ewa inertia: 107464.300085 
Minibatch iteration 189/16100: mean batch inertia: 106636.095740, ewa inertia: 107453.961533 
Minibatch iteration 190/16100: mean batch inertia: 106668.333199, ewa inertia: 107444.154463 
Minibatch iteration 191/16100: mean batch inertia: 106732.683060, ewa inertia: 107435.273100 
Minibatch iteration 192/16100: mean batch inertia: 107072.200524, ewa inertia: 107430.740831 
Minibatch iteration 193/16100: mean batch inertia: 107367.395103, ewa inertia: 107429.950081 
Minibatch iteration 194/16100: mean batch inertia: 107202.011031, ewa inertia: 107427.104696 
Minibatch iteration 195/16100: mean batch inertia: 105122.243989, ewa inertia: 107398.332908 
Minibatch iteration 196/16100: mean batch inertia: 106856.179710, ewa inertia: 107391.565160 
Minibatch iteration 197/16100: mean batch inertia: 107009.928206, ewa inertia: 107386.801150 
Minibatch iteration 198/16100: mean batch inertia: 106715.351097, ewa inertia: 107378.419378 
Minibatch iteration 199/16100: mean batch inertia: 105536.370060, ewa inertia: 107355.424906 
Minibatch iteration 200/16100: mean batch inertia: 106727.936925, ewa inertia: 107347.591916 
Minibatch iteration 201/16100: mean batch inertia: 107002.406327, ewa inertia: 107343.282933 
Minibatch iteration 202/16100: mean batch inertia: 106330.836794, ewa inertia: 107330.644474 
Minibatch iteration 203/16100: mean batch inertia: 106261.619101, ewa inertia: 107317.299732 
Minibatch iteration 204/16100: mean batch inertia: 106742.834856, ewa inertia: 107310.128634 
Minibatch iteration 205/16100: mean batch inertia: 107105.762387, ewa inertia: 107307.577511 
Minibatch iteration 206/16100: mean batch inertia: 106788.584827, ewa inertia: 107301.098878 
Minibatch iteration 207/16100: mean batch inertia: 106686.880624, ewa inertia: 107293.431535 
Minibatch iteration 208/16100: mean batch inertia: 105396.080329, ewa inertia: 107269.746725 
Minibatch iteration 209/16100: mean batch inertia: 105522.870331, ewa inertia: 107247.940306 
Minibatch iteration 210/16100: mean batch inertia: 108026.090720, ewa inertia: 107257.654029 
Minibatch iteration 211/16100: mean batch inertia: 106857.737964, ewa inertia: 107252.661840 
Minibatch iteration 212/16100: mean batch inertia: 106170.112944, ewa inertia: 107239.148283 
Minibatch iteration 213/16100: mean batch inertia: 105942.474624, ewa inertia: 107222.961786 
Minibatch iteration 214/16100: mean batch inertia: 105895.256116, ewa inertia: 107206.387914 
Minibatch iteration 215/16100: mean batch inertia: 106559.843067, ewa inertia: 107198.317035 
Minibatch iteration 216/16100: mean batch inertia: 107246.422658, ewa inertia: 107198.917542 
Minibatch iteration 217/16100: mean batch inertia: 106804.394606, ewa inertia: 107193.992676 
Minibatch iteration 218/16100: mean batch inertia: 106486.454114, ewa inertia: 107185.160407 
Minibatch iteration 219/16100: mean batch inertia: 104863.942086, ewa inertia: 107156.184425 
Minibatch iteration 220/16100: mean batch inertia: 107044.589043, ewa inertia: 107154.791369 
Minibatch iteration 221/16100: mean batch inertia: 106282.972867, ewa inertia: 107143.908379 
Minibatch iteration 222/16100: mean batch inertia: 106615.700985, ewa inertia: 107137.314717 
Minibatch iteration 223/16100: mean batch inertia: 107074.673759, ewa inertia: 107136.532764 
Minibatch iteration 224/16100: mean batch inertia: 106866.989377, ewa inertia: 107133.168029 
Minibatch iteration 225/16100: mean batch inertia: 107497.812342, ewa inertia: 107137.719918 
Minibatch iteration 226/16100: mean batch inertia: 107111.683011, ewa inertia: 107137.394897 
Minibatch iteration 227/16100: mean batch inertia: 107102.205610, ewa inertia: 107136.955626 
Minibatch iteration 228/16100: mean batch inertia: 107495.415617, ewa inertia: 107141.430315 
Minibatch iteration 229/16100: mean batch inertia: 106032.203935, ewa inertia: 107127.583740 
Minibatch iteration 230/16100: mean batch inertia: 107032.676650, ewa inertia: 107126.399006 
Minibatch iteration 231/16100: mean batch inertia: 106413.784761, ewa inertia: 107117.503376 
Minibatch iteration 232/16100: mean batch inertia: 106397.665961, ewa inertia: 107108.517580 
Minibatch iteration 233/16100: mean batch inertia: 107380.854236, ewa inertia: 107111.917183 
Minibatch iteration 234/16100: mean batch inertia: 107100.862229, ewa inertia: 107111.779183 
Minibatch iteration 235/16100: mean batch inertia: 106967.045325, ewa inertia: 107109.972457 
Minibatch iteration 236/16100: mean batch inertia: 107010.954320, ewa inertia: 107108.736405 
Minibatch iteration 237/16100: mean batch inertia: 106017.746277, ewa inertia: 107095.117474 
Minibatch iteration 238/16100: mean batch inertia: 107777.682898, ewa inertia: 107103.638001 
Minibatch iteration 239/16100: mean batch inertia: 107184.475980, ewa inertia: 107104.647109 
Minibatch iteration 240/16100: mean batch inertia: 106318.199208, ewa inertia: 107094.829808 
Minibatch iteration 241/16100: mean batch inertia: 106824.724141, ewa inertia: 107091.458054 
Minibatch iteration 242/16100: mean batch inertia: 107149.742563, ewa inertia: 107092.185625 
Minibatch iteration 243/16100: mean batch inertia: 106649.016339, ewa inertia: 107086.653502 
Minibatch iteration 244/16100: mean batch inertia: 106684.187429, ewa inertia: 107081.629481 
Minibatch iteration 245/16100: mean batch inertia: 106346.545726, ewa inertia: 107072.453363 
Minibatch iteration 246/16100: mean batch inertia: 106413.960755, ewa inertia: 107064.233339 
Minibatch iteration 247/16100: mean batch inertia: 106647.535957, ewa inertia: 107059.031667 
Minibatch iteration 248/16100: mean batch inertia: 106040.988980, ewa inertia: 107046.323346 
Minibatch iteration 249/16100: mean batch inertia: 107473.159454, ewa inertia: 107051.651581 
Minibatch iteration 250/16100: mean batch inertia: 106390.363319, ewa inertia: 107043.396658 
Minibatch iteration 251/16100: mean batch inertia: 106970.189615, ewa inertia: 107042.482808 
Minibatch iteration 252/16100: mean batch inertia: 106568.826189, ewa inertia: 107036.570109 
Minibatch iteration 253/16100: mean batch inertia: 107502.289213, ewa inertia: 107042.383723 
Minibatch iteration 254/16100: mean batch inertia: 106342.522778, ewa inertia: 107033.647295 
Minibatch iteration 255/16100: mean batch inertia: 106386.191434, ewa inertia: 107025.565044 
Minibatch iteration 256/16100: mean batch inertia: 107017.038179, ewa inertia: 107025.458602 
Minibatch iteration 257/16100: mean batch inertia: 107105.632737, ewa inertia: 107026.459423 
Minibatch iteration 258/16100: mean batch inertia: 106435.177550, ewa inertia: 107019.078397 
Minibatch iteration 259/16100: mean batch inertia: 106811.168151, ewa inertia: 107016.483034 
Minibatch iteration 260/16100: mean batch inertia: 106694.713264, ewa inertia: 107012.466353 
Minibatch iteration 261/16100: mean batch inertia: 106546.824305, ewa inertia: 107006.653700 
Minibatch iteration 262/16100: mean batch inertia: 106267.791393, ewa inertia: 106997.430414 
Minibatch iteration 263/16100: mean batch inertia: 107540.658508, ewa inertia: 107004.211580 
Minibatch iteration 264/16100: mean batch inertia: 107168.320092, ewa inertia: 107006.260162 
Minibatch iteration 265/16100: mean batch inertia: 106501.275699, ewa inertia: 106999.956394 
Minibatch iteration 266/16100: mean batch inertia: 107473.650138, ewa inertia: 107005.869557 
Minibatch iteration 267/16100: mean batch inertia: 106893.237364, ewa inertia: 107004.463559 
Minibatch iteration 268/16100: mean batch inertia: 106607.432293, ewa inertia: 106999.507381 
Minibatch iteration 269/16100: mean batch inertia: 106788.275486, ewa inertia: 106996.870554 
Minibatch iteration 270/16100: mean batch inertia: 106988.084110, ewa inertia: 106996.760872 
Minibatch iteration 271/16100: mean batch inertia: 107153.976646, ewa inertia: 106998.723411 
Minibatch iteration 272/16100: mean batch inertia: 107281.203503, ewa inertia: 107002.249636 
Minibatch iteration 273/16100: mean batch inertia: 107645.820150, ewa inertia: 107010.283386 
Minibatch iteration 274/16100: mean batch inertia: 105834.064890, ewa inertia: 106995.600542 
Minibatch iteration 275/16100: mean batch inertia: 107477.081802, ewa inertia: 107001.610917 
Minibatch iteration 276/16100: mean batch inertia: 107905.483447, ewa inertia: 107012.894041 
Minibatch iteration 277/16100: mean batch inertia: 107175.853220, ewa inertia: 107014.928275 
Minibatch iteration 278/16100: mean batch inertia: 107141.125637, ewa inertia: 107016.503609 
Minibatch iteration 279/16100: mean batch inertia: 107126.612698, ewa inertia: 107017.878111 
Minibatch iteration 280/16100: mean batch inertia: 105862.170164, ewa inertia: 107003.451302 
Minibatch iteration 281/16100: mean batch inertia: 106488.277986, ewa inertia: 106997.020346 
Minibatch iteration 282/16100: mean batch inertia: 106461.821252, ewa inertia: 106990.339406 
Minibatch iteration 283/16100: mean batch inertia: 106464.736203, ewa inertia: 106983.778253 
Minibatch iteration 284/16100: mean batch inertia: 105446.417702, ewa inertia: 106964.587240 
Minibatch iteration 285/16100: mean batch inertia: 106994.129328, ewa inertia: 106964.956016 
Minibatch iteration 286/16100: mean batch inertia: 106373.586240, ewa inertia: 106957.573893 
Minibatch iteration 287/16100: mean batch inertia: 107256.435014, ewa inertia: 106961.304604 
Minibatch iteration 288/16100: mean batch inertia: 107518.317362, ewa inertia: 106968.257845 
Minibatch iteration 289/16100: mean batch inertia: 107064.853646, ewa inertia: 106969.463660 
Minibatch iteration 290/16100: mean batch inertia: 107043.837999, ewa inertia: 106970.392081 
Minibatch iteration 291/16100: mean batch inertia: 106142.188603, ewa inertia: 106960.053541 
Minibatch iteration 292/16100: mean batch inertia: 107001.014650, ewa inertia: 106960.564862 
Minibatch iteration 293/16100: mean batch inertia: 106359.944611, ewa inertia: 106953.067265 
Minibatch iteration 294/16100: mean batch inertia: 105853.227387, ewa inertia: 106939.337862 
Minibatch iteration 295/16100: mean batch inertia: 108041.692045, ewa inertia: 106953.098651 
Minibatch iteration 296/16100: mean batch inertia: 106966.889643, ewa inertia: 106953.270805 
Minibatch iteration 297/16100: mean batch inertia: 107219.564858, ewa inertia: 106956.594978 
Minibatch iteration 298/16100: mean batch inertia: 107229.300322, ewa inertia: 106959.999184 
Minibatch iteration 299/16100: mean batch inertia: 107371.027034, ewa inertia: 106965.130083 
Minibatch iteration 300/16100: mean batch inertia: 107755.053458, ewa inertia: 106974.990769 
Minibatch iteration 301/16100: mean batch inertia: 106485.129309, ewa inertia: 106968.875783 
Minibatch iteration 302/16100: mean batch inertia: 107398.536248, ewa inertia: 106974.239274 
Minibatch iteration 303/16100: mean batch inertia: 106024.745196, ewa inertia: 106962.386652 
Minibatch iteration 304/16100: mean batch inertia: 106841.172053, ewa inertia: 106960.873519 
Converged (lack of improvement in inertia) at iteration 304/16100
Computing label assignment and total inertia

Creation des histogrammes

Nous créons ici pour chaque image, un histogramme.Pour chaque image on aura un vecteur de k valeur. Pour chaque keypoints dans une image, nous trouverons son point centre le plus proche.Enfin, on augmentera de 1 sa valeur afin de normaliser.

kmeans.verbose = False

histo_list = []

for img in os.listdir(datadir):
    img_array=imread(os.path.join(datadir,img))
    gray_img = cv2.cvtColor(img_array, cv2.COLOR_BGR2GRAY)
    kp, des = orb.detectAndCompute(gray_img, None)

    histo = np.zeros(k)
    nkp = np.size(kp)

    for d in des:
        idx = kmeans.predict([d])
        histo[idx] += 1/nkp # Because we need normalized histograms, we add 1/nkp directly

    histo_list.append(histo)

C:\Users\malik\anaconda3\lib\site-packages\PIL\Image.py:2834: DecompressionBombWarning: Image size (93680328 pixels) exceeds limit of 89478485 pixels, could be decompression bomb DOS attack.
  warnings.warn(

plt.hist(histo_list[0])
len(histo_list)

1049

Train and Split

X = np.array(histo_list)
Y = df['categories_niv1']

from sklearn.model_selection import train_test_split
x_train,x_test,y_train,y_test=train_test_split(X,Y,test_size=0.33,random_state=77)

Model algorithme XGBOOST de classification

mod = XGBRFClassifier()
mod.fit(x_train,y_train)
print(f"The model Xgboost is {accuracy_score(mod.predict(x_test),y_test)*100}% accurate")

C:\Users\malik\anaconda3\lib\site-packages\xgboost\sklearn.py:888: UserWarning: The use of label encoder in XGBClassifier is deprecated and will be removed in a future release. To remove this warning, do the following: 1) Pass option use_label_encoder=False when constructing XGBClassifier object; and 2) Encode your labels (y) as integers starting with 0, i.e. 0, 1, 2, ..., [num_class - 1].
  warnings.warn(label_encoder_deprecation_msg, UserWarning)

[23:56:35] WARNING: ..\src\learner.cc:1061: Starting in XGBoost 1.3.0, the default evaluation metric used with the objective 'multi:softprob' was changed from 'merror' to 'mlogloss'. Explicitly set eval_metric if you'd like to restore the old behavior.
The model Xgboost is 16.714697406340058% accurate

Reduction et Visualisation

from sklearn.decomposition import PCA
from sklearn.manifold import TSNE
import matplotlib.pyplot as plt
import seaborn as sns
def reduction_visualisation(method):
    if(method == "PCA"):
        mod = PCA(n_components=2)
        title = "PCA decomposition"  # for the plot
    elif(method == "TSNE"):
        mod = TSNE(n_components=2)
        title = "TSNE decomposition"  # for the plot
    # Fit and transform the features
    principal_components = mod.fit_transform(X)

    # Put them into a dataframe
    df_features = pd.DataFrame(data=principal_components,
                     columns=['PC1', 'PC2'])

    # Now we have to paste each row's label and its meaning
    # Convert labels array to df
    df.replace([0,1,2,3,4,5,6],['Baby Care','Beauty and Personal Care','Computers','Home Decor & Festive Needs','Home Furnishing','Kitchen & Dining','Watches'],inplace=True)
    df_labels = df['categories_niv1']
    df_full = pd.concat([df_features, df_labels], axis=1)


    # Plot
    plt.figure(figsize=(10,10))
    sns.scatterplot(x='PC1',
                    y='PC2',
                    hue="categories_niv1",
                    data=df_full,
                    palette=["red", "pink", "royalblue", "greenyellow", "lightseagreen","black","purple"],
                    alpha=.7).set_title(title);

reduction_visualisation('PCA')
reduction_visualisation('TSNE')

Conclusion -ImageClassification = non

3 - Classification sur description avec NLTK

nltk.download('punkt')
nltk.download('stopwords')
stemmer = PorterStemmer()
dataset = dataWCL.copy()
data = []

for i in range(dataset.shape[0]):
    description = dataset.iloc[i, 1]
    # remove non alphabatic characters
    description = re.sub('[^A-Za-z]', ' ', description)
    # make words lowercase, because Example and example will be considered as two words
    description = description.lower()
    # tokenising
    tokenized_description = wt(description)
    # remove stop words and stemming
    text_processed = []
    for word in tokenized_description:
        if word not in set(stopwords.words('english')):
            text_processed.append(stemmer.stem(word))
    description_text = " ".join(text_processed)
    data.append(description_text)

# creating the feature matrix 
matrix = CountVectorizer(max_features=1000)
X = matrix.fit_transform(data).toarray()
y = dataset.iloc[:, 0]

# split train and test data
X_train, X_test, y_train, y_test = train_test_split(X, y)

# Naive Bayes 
classifier = GaussianNB()
classifier.fit(X_train, y_train)

# predict class
y_pred = classifier.predict(X_test)

# Confusion matrix
cm = confusion_matrix(y_test, y_pred)
cr = classification_report(y_test, y_pred)

accuracy = accuracy_score(y_test, y_pred)
print("GaussianNB accuracy score is: ",accuracy)

[nltk_data] Downloading package punkt to
[nltk_data]     C:\Users\malik\AppData\Roaming\nltk_data...
[nltk_data]   Package punkt is already up-to-date!
[nltk_data] Downloading package stopwords to
[nltk_data]     C:\Users\malik\AppData\Roaming\nltk_data...
[nltk_data]   Package stopwords is already up-to-date!

GaussianNB accuracy score is:  0.8479087452471483

#from sklearn.metrics import plot_confusion_matrix
#plot_confusion_matrix(classifier, X_test, y_test)
from sklearn.metrics import ConfusionMatrixDisplay
disp = ConfusionMatrixDisplay(confusion_matrix=cm,display_labels=classifier.classes_)
disp.plot()
plt.show()

# Classification report
print(cr)

                            precision    recall  f1-score   support

                 Baby Care       0.74      0.64      0.69        36
  Beauty and Personal Care       0.74      0.82      0.78        34
                 Computers       0.89      0.89      0.89        37
Home Decor & Festive Needs       0.91      0.78      0.84        40
           Home Furnishing       0.85      0.93      0.89        42
          Kitchen & Dining       0.90      0.93      0.91        40
                   Watches       0.89      0.94      0.91        34

                  accuracy                           0.85       263
                 macro avg       0.85      0.85      0.84       263
              weighted avg       0.85      0.85      0.85       263

Reduction et Visualisation

reduction_visualisation('PCA')
reduction_visualisation('TSNE')

Conclusion - TextClassification = oui