Skip to content

Latest commit

 

History

History
23 lines (18 loc) · 1.4 KB

README.md

File metadata and controls

23 lines (18 loc) · 1.4 KB
Raw

nlp

NLP practice projects and examples. All data files in .gitignore.

Example projects: https://medium.com/coders-camp/20-machine-learning-projects-on-nlp-582effe73b9c

ML on M1: https://caffeinedev.medium.com/how-to-install-tensorflow-on-m1-mac-8e9b91d93706

For Keras/TensorFlow on an Apple M1, I'm using conda miniforge with TensorFlow installed in a virtual environment (conda activate mlp, after install). Much of the NLP uses NLTK rather than spaCy, with separately downloaded Glove word embeddings.

Webapp: https://wordchef.app/topic_modeling

  1. Binary sentence classification via CNN/RNN on Biden/Trump tweets.
  2. Text summarization via Glove embeddings, cosine similarity matrix, and pagerank.
  3. Sentiment analysis via logistic regression.
  4. Keyword extraction via TF-IDF.
  5. Hate speech detection using SGDClassifier.
  6. Next word prediction using Keras/TF.
  7. Spam detection using SMS data, CountVectorizer (bag-of-words), and Multinomial Naive Bayes.
  8. Text classification, part 2 using TF/Keras, IMDB reviews for positive/negative sentiment.
  9. Spelling correction using TextBlob (quick).
  10. Named entity recognition via LSTM. Too long to train on M1.
  11. Twitter sentiment analysis, part II. Use Naive Bayes sentiment classifier trained on positive/negative labeled tweets.
  12. Topic modeling. Use doc2vec embeddings, k-means clustering, TF-IDF scores to label clusters with topic.