NLP using random forest classification

Author: Nikronic

Part 7 - Natural Language Processing/Section 20 - Natural Language Processing/nlp_random-forest.py

@@ -0,0 +1,76 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Tue Sep 18 17:56:22 2018
+
+@author: Mohammad Doosti Lakhani
+"""
+
+# -*- coding: utf-8 -*-
+"""
+Created on Mon Sep 17 21:44:06 2018
+
+@author: Mohammad Doosti Lakhani
+"""
+
+# Importing the libraries
+import pandas as pd
+
+# Importing the dataset
+dataset = pd.read_csv('Restaurant_Reviews.tsv', delimiter = '\t', quoting = 3) # remove quoting marks
+x = dataset.iloc[:,0].values
+y = dataset.iloc[:,1].values
+
+# Cleaning the Text
+import re
+from nltk.corpus import stopwords # removing useless words
+from nltk.stem.porter import PorterStemmer # getting root of words
+stopwords_list = stopwords.words('english')
+ps = PorterStemmer()
+
+corpus = []
+for i in range(0,len(x)):
+    corp = re.sub(pattern = '[^a-zA-Z]', repl = ' ', string = x[i])
+    corp = corp.split()
+    corp = [ps.stem(word) for word in corp]
+    corp = ' '.join(corp)
+    corpus.append(corp)
+
+# Creating the Bag of Words model
+from sklearn.feature_extraction.text import CountVectorizer # text to column vector os sparse matrix
+cv = CountVectorizer(lowercase=True, stop_words = stopwords_list) # ignoring useless words and to lowercase
+x_sparse = cv.fit_transform(corpus) # bag of word sparse matrix
+x = x_sparse.todense() # Convert sparse matrix to dense matrix
+
+# Splitting dataset into Train set and Test set
+from sklearn.model_selection import train_test_split
+x_train,x_test,y_train,y_test = train_test_split(x,y, train_size = 0.85 , random_state=8)
+
+# Fitting the Naive Bayes (Gausiian form) model to the train set
+from sklearn.ensemble import RandomForestClassifier
+classifier = RandomForestClassifier(n_estimators=12,criterion='entropy',random_state=0)
+classifier = classifier.fit(x_train,y_train)
+
+# Make the prediction on train set
+y_train_pred = classifier.predict(x_train)
+
+# Make the prediction on train set
+y_test_pred = classifier.predict(x_test)
+
+# Acurracy on test and train set
+from sklearn.metrics import confusion_matrix
+cm_train = confusion_matrix(y_train,y_train_pred)
+cm_test = confusion_matrix(y_test,y_test_pred)
+
+import os
+import sys
+
+scriptpath = "../../Tools" # functions of acc
+# Add the directory containing your module to the Python path
+sys.path.append(os.path.abspath(scriptpath))
+import accuracy as ac
+
+t_train,f_train,acc_train = ac.accuracy_on_cm(cm_train)
+print('Train status = #{} True, #{} False, %{} Accuracy'.format(t_train,f_train,acc_train*100))
+
+t_test,f_test,acc_test = ac.accuracy_on_cm(cm_test)
+print('Test status = #{} True, #{} False, %{} Accuracy'.format(t_test,f_test,acc_test*100))