NLP Projects

Project Description

• TFIDF for Reuters Articles in XML

  • Extracted titles and paragraphs from Reuters articles in the XML format using ElementTree
  • Tokenized and stemmed texts with NLTK, and determined TFIDF of the most common words using TfidfVectorizer

• Sentiment Analysis with Naive Bayes using PySpark

  • Performed data cleaning and transformation, and estimate TF using PySpark RDD
  • Built a Naive Bayes model to perform sentiment analysis and achieved an accuracy of 82.5%

• Sentiment Analysis of Tweets

  • Parsed and stemmed tweet texts, and determined TF of the most common words
  • Classified the tweet sentiment using the regularized logistic regression, LDA, and KNN