News category prediction using Natural language processing [NLP]

In this tutorial, we will work on the news articles dataset and categorize the articles based on the content. So let’s learn how to predict news category using NLP (Natural Language Processing) with Python.

The dataset we will be using is:

News category prediction using NLP Dataset zip file – Download Dataset

News category prediction using NLP in Python using scikit-learn

First, we will start by importing the required libraries:

%matplotlib inline
import re
import matplotlib
import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB
from sklearn.metrics import accuracy_score
from sklearn.multiclass import OneVsRestClassifier
from nltk.corpus import stopwords
stop_words = set(stopwords.words('english'))
from sklearn.svm import LinearSVC
from nltk.stem.porter import PorterStemmer
from nltk.stem import WordNetLemmatizer,LancasterStemmer 
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import Pipeline
import seaborn as sns
from keras import utils as np_utils
from keras.preprocessing import sequence
from keras.preprocessing.text import Tokenizer
from keras.models import Sequential,Model
from keras.layers import Dense, Dropout, Activation
from keras.layers import Embedding,Input,LSTM
from keras.layers import Conv1D, GlobalMaxPooling1D
import tensorflow as tf
from sklearn.externals import joblib
from textblob import TextBlob
from keras.optimizers import RMSprop,Adam
from keras.callbacks import EarlyStopping

Importing the dataset

df = pd.read_excel("Data_Train.xlsx")

The article may contain many repetitive words like a the, and many other prepositions, connectors. Because these words repeat very often, we need to remove these words. For this, we write a function to clean the articles. Cleaning includes removal of punctuation marks, stop words. The text is converted to the lowercase so that there is no confusion among the uppercase and lowercase words.  Lemmatization that involves grouping together different inflected forms of word so that they can be analyzed as a single term. It also involves removing the apostrophes.

def clean_text(text):
    text = text.lower()
    text = re.sub(r"what's", "what is ", text)
    text = re.sub(r"\'s", " ", text)
    text = re.sub(r"\'ve", " have ", text)
    text = re.sub(r"can't", "can not ", text)
    text = re.sub(r"n't", " not ", text)
    text = re.sub(r"i'm", "i am ", text)
    text = re.sub(r"\'re", " are ", text)
    text = re.sub(r"\'d", " would ", text)
    text = re.sub(r"\'ll", " will ", text)
    text = re.sub(r"\'scuse", " excuse ", text)
    text = re.sub('\W', ' ', text)
    text = re.sub('\s+', ' ', text)
    text = text.strip(' ')
    text=re.sub('[^a-zA-Z]',' ',text)
    text=text.lower()
    text=text.split()
    lemmatizer = WordNetLemmatizer() 
    text=[lemmatizer.lemmatize(word) for word in text if not word in set(stopwords.words('english'))]
    text=' '.join(text)

    return text

We now apply this method to the text in the data frame in order to get the relevant information.

df['STORY']=df['STORY'].map(lambda story:clean_text(story))

Now, we will split the dataset into training and test sets so that we can train the model and validate it on the test set:

train, test = train_test_split(df, random_state=42, test_size=0.2)
x_train = train.STORY
x_test = test.STORY
y_train=train.SECTION
y_test=test.SECTION

After splitting, we create a matrix based on the frequency of the word in the content.

vectorizer=TfidfVectorizer(max_df=0.9,min_df=1,stop_words='english')
train_vectors=vectorizer.fit_transform(x_train)
test_vectors=vectorizer.transform(x_test)
total_vectors=vectorizer.transform(df['STORY'])

Creating a classifier to categorize the articles:

from sklearn.neural_network import MLPClassifier
mlp=MLPClassifier()
mlp.fit(train_vectors,y_train)
mlp_prediction=mlp.predict(test_vectors)
accuracy_score(y_test, mlp_prediction)

Output: It shows the accuracy of the model

from sklearn.neural_network import MLPClassifier
mlp=MLPClassifier()

mlp.fit(train_vectors,y_train)
Out[8]: 
MLPClassifier(activation='relu', alpha=0.0001, batch_size='auto', beta_1=0.9,
              beta_2=0.999, early_stopping=False, epsilon=1e-08,
              hidden_layer_sizes=(100,), learning_rate='constant',
              learning_rate_init=0.001, max_iter=200, momentum=0.9,
              n_iter_no_change=10, nesterovs_momentum=True, power_t=0.5,
              random_state=None, shuffle=True, solver='adam', tol=0.0001,
              validation_fraction=0.1, verbose=False, warm_start=False)

mlp_prediction=mlp.predict(test_vectors)
accuracy_score(y_test, mlp_prediction)

Out[9]: 0.9796854521625163