Comprehend is an AWS service that can perform many NLP tasks like translation, topic modeling, sentiment anlaysis, etc...
keywords: AWS, Comprehend, SageMaker, sentiment analysis
Sentiment analysis is the NLP task of understanding the attitudes or emotions in a piece of text. This is often used to analyze customers' feedback in social media or eCommerce websites on products. Here I will appy sentiment analysis on customers' reviews on Amazon products using a range of classifiers as well as deep learning model.
keywords: Bag of Words, Naive Bayes classifier, SVM, KNN, XGBoosting classifier, Keras, LSTM
Using LSTM and attention mechanism, and the amazon fine food reviews from kaggle, a text summarizer is finally born to save us from long exhausting reviews.
keywords: LSTM, seq2seq, Bahdanau attention, kaggle dataset, food reviews, GloVe
fasttext is a new facebook deeplearning framework that uses light learning algorithms to train models fastly. This text classifier takes the news title and predicts whether the news is fake or real.
keywords: fasttext, fake/real news
The aim of this project is to apply spell checking & correction as well as other NLP tasks like regex search. So I used PyQt5 to build a simple texteditor and integrated the spell checking & correction functionality in it. This is a non-DL NLP project, No fancy networks used, just a decent set of real words, a language model, an error model, a candidate model, and shokran!
keywords: text editor, levenshtien distance, PyQt5, spell checking/correction, Regex search