seroetr / Sentiment_Analysis_using_Machine_Learning_and_Deep_Learning

Application of Machine and Deep Learning to Sentiment Analysis

Machine Learning Algorithms and Fine Tuning

• Machine Learning Algorithms and Fine Tuning are performed by using pipeline and hyperparameter search for fine tuning as an example.

Twitter sentiment analysis is performed by using dimensionality reduction techniques in ML algorithms

• Application of ML algorithms is performed by using pipeline and hyperparameter search for fine tuning as an example.
• Detailed information can be found via course link: https://bit.ly/intro_nlp

IMDB Movie Reviews Sentiment Analysis with TF-IDF

• Using pipeline with the GridSearchCV to fine-tune hyperparameters, Logistic regression and SVM model are used to predict the sentiment analysis of a movie review. After that, model is saved and loaded again to test it with text.
• Pickle is used to save and load the model.

Sentiment Analysis with Deep Learning using ANN and CNN in Tensorflow

• Simple Artificial Neural Network (ANN) and 1D Convolutional Neural Network (CNN) models are used to predict sentiment analysis for IMDB movei reviews, which is a binary classification problem.
• Spacy en_core_web_lg pretrained model (https://spacy.io/models/en) is used to convert texts to vectors in dataframes.
• Feature scaling is applied to X_train and X_test using MinMaxScaler before being fed inputs into the neural network.