aniksarakash / advanced-tourist-guide-system

A web based app designed and developed using Django framework of Python with the power of Deep learning for sentiment analysis. Aim of the project is to provide travelers a solution to make their travels more easy & convenient.

Advanced-Tourist-Guide-System

A web based app designed and developed using Django framework of Python with the power of Deep learning for sentiment analysis. Aim of the project is to provide travelers a solution to make their travels more easy & convenient.

• A web based application designed and developed using Django framework of python. Aim of the project is to make a tool using which, users can get information about the places very easily.

Features of the project : - User can search the place via, place name as well as via location name. - Home screen of the portal will display the top most rated places from the database. - User will also get the similar places & also other places nearby his search. - Reviews will be stored as a comment for the place and also, sentiment analysis will be done using Deep Learning and ratings will be stored. - User can also give feedback to the system - User can get direction from his current location to the place.

TOOLS AND TECHNOLOGIES USED : - Frontend : HTML, JS, CSS - Backend : Python, DJango framework - Sentiment Analysis : CNN-LSTM-MAXPOOL

DATASET FOR MODEL TRAINING : - GoogleNews-vectors-negative300.bin.gz https://s3.amazonaws.com/dl4j-distribution/GoogleNews-vectors-negative300.bin.gz

- Sentiment-Analysis-Dataset.zip
	http://thinknook.com/wp-content/uploads/2012/09/Sentiment-Analysis-Dataset.zip

ALGORITHM USED : - CNN with LSTM & MAXPOOL

METHODOLOGY : - Firstly, we downloaded the datasets from the above given links.

1) importing the required modules :

	import gensim.models.keyedvectors as word2vec 
	from nltk.tokenize import RegexpTokenizer
	import keras
	import numpy as np
	import pandas as pd
	import matplotlib.pyplot as plt
	import sklearn



2) PreProcessing  : 

- read CSV file containing tweets and labels, using Pandas , to get a dataframe : 
	tweetsData = pd.read_csv('Sentiment Analysis Dataset.csv', skiprows=[8835],nrows=40000) 

- Dividing the dataset into features and lables :
	tweets = tweetsData['SentimentText']
	labels = tweetsData['Sentiment']

- Lower and split the dialog and use regular expression to keep only letters we will use nltk Regular expression package :
	tkr = RegexpTokenizer('[a-zA-Z@]+')

- Use pretrained Word2Vec model from google but trim the word list to 50,000 compared to 3,000,000 in the originaL Google pretrained model :
	w2vModel = word2vec.KeyedVectors.load_word2vec_format('GoogleNews-vectors-negative300.bin')

- Convert words to integers :
	tokenizer = Tokenizer(num_words=len(tweets_split)+1)
	tokenizer.fit_on_texts(tweets_split)
	X = tokenizer.texts_to_sequences(tweets_split)

- create a embedding layer using Google pre triained word2vec (50000 words) :
	embedding_layer = Embedding(input_dim, output_dim, weights, input_length)



3) Building Model Architecture :

	- Create model combining LSTM with 1D Convonutional layer and MaxPool layer

		model = Sequential()
		model.add(embedding_layer)
		model.add(Conv1D(filters=64, kernel_size=5, activation='relu', padding='causal'))
		model.add(MaxPooling1D(pool_size=2))
		model.add(Dropout(0.7))
		model.add(LSTM(units=lstm_out))
		model.add(Dropout(0.7))
		model.add(Dense(1, activation='sigmoid'))
		model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])




4) Fitting training data to model :
	- split dataset
		X_train, X_test, Y_train, Y_test = train_test_split(X, labels, test_size= 0.1)

	- fit model
		batch_size = 32
		training = model.fit(X_train, Y_train, epochs=5, verbose=1, batch_size=batch_size)



5) Evaluatiing Model :

	- analyze the results :
		score, acc = model.evaluate(X_test, Y_test, verbose = 2, batch_size=batch_size)
		y_pred = model.predict(X_test)

	- Other accuracy metrices :
		y_pred = (y_pred > 0.5)

	- confusion metrix :
		cm = confusion_matrix(Y_test, y_pred)
		print(cm)




6) Visualizing the model :
	- ROC AUC curve
		rocAuc = roc_auc_score(Y_test, y_pred)

		falsePositiveRate, truePositiveRate, _ = roc_curve(Y_test, y_pred)

		plt.figure()

		plt.plot(falsePositiveRate, truePositiveRate, color='green')
		plt.plot([0, 1], [0, 1], color='navy', lw=3, linestyle='--')
		plt.xlim([0.0, 1.0])
		plt.ylim([0.0, 1.05])
		plt.xlabel('False Positive Rate')
		plt.ylabel('True Positive Rate')
		plt.title('Receiver Operating Characteristic of Sentiiment Analysis Model')
		plt.legend(loc="lower right")
		plt.show()