Sentimentanalysis_CNN

Sentiment Analysis done using CNN (Convolutional Neural Net)

Pre-requistics

Python- 3.7
Torch - 1.0.0
numpy -1.16.4
sklearn - 0.21.3
nltk - 3.4.4
pandas - 0.25.0

Files

1. data.csv - Contains the dataset having the sentence and its sentiment label.
2. Text_cleaner.py - This file will clean the dataset by removing the stopwords, punctuations and converting the format to UTF-8
3. Conv_preprocessing.py - This file contains the main code that needs to be run to predict the sentiment.
4. ConvClassifier.py - This file contains the training, testing and building a CNN model required for the sentiment analysis.

Note:- Text cleaner file is made as a separate file for re-usability and saving time to avoid running the text cleaning process for each and every time Pre_processing.py file is run.

Downloads and Setup

Once you clone this repo, run the Conv_preprocessing.py file to do the sentiment analysis and to train the model.

Evalution metric

Evalution metric used here is the accuracy.