Sentiment Analysis

This project analyses the performance of different methods in the challenging task of sentiment analysis. In order to do so, Amazon Kindle data was obtained from (https://nijianmo.github.io/amazon/index.html).

Implementation

The implementation process consisted in four main steps:

• Pre-processing: Sampling (to overcome the class imbalance issue) and text cleaning.
• Exploratory Data Analysis (EDA): N-grams, word count, etc.
• Classification: Used both classical machine learning methods (such as Gradient Boosting, SVMs and Gaussian Naive Bayes) and deep learning methods (e.g. BERT, RoBERTa, LSTMs)
• Text generation

Enviroment

• Python 3.6+
• Pandas
• Spacy
• Tqdm
• Spacymoji
• Numpy
• Sklearn
• Imblearn
• Matplotlib
• Seaborn
• Gensim
• PyLDAvis
• Logging
• Nltk
• Wordcloud
• Torchbearer
• PyTorch
• Transformers

Install all python modules with

pip install -r requirements.txt

or if you have different versions of Python installed:

pip3 install -r requirements.txt

Authors

• Alex Newton - xandernewton

• Diogo Pereira - DiogoP98