gammaraysky/nltk_exercises

small compilation of classification and clustering problems

nltk exercises

small compilation of classification/clustering/topic modelling problems

• nltk_bookreview_sentimentanalysis.ipynb
  • positive/negative sentiment analysis on amazon book reviews
• nltk_spam_classifier.ipynb
  • positive/negative spam prediction on sms text messages
• nltk_bbcnews_topicmodelling.ipynb
  • topic modelling of 2000 BBC news articles into categories.
    • preprocess using CountVectorizer and tf-idf.
    • apply PCA and k-means to estimate optimal n_topics
    • Clustering using t-SNE and UMAP reductions
    • LDA topic modelling with gensim vs sklearn
    • wordcloud visualisation of topic keywords
    • BERTopic topic modelling, visualize topic distance, hierarchical linkage