sourav-raj / nlp-projects Goto Github PK

financial-sentiment-analysis

Perform pre-processing steps like Removing Punctuations, Numbers, and Special Characters, Stop Words in dataset.
Normalize Sentence column by using Stemming or Lemmatization. Plot word cloud for the sentence. Perform label encoding technique for sentiment column. Create a Word2Vec object , and train a MultinomialNB model. Create a CBoW object and again train a MultinomialNB model. Compare Word2Vec and CBoW.

trip-advisor-reviews

use pre-trained embedding layer ( https://tfhub.dev/google/nnlm-en-dim128/2 from TensorFlow hub for this Sequential Model layers - Use AT LEAST 5 hidden layers, Add L1 regularization to all the layers. Hyperparameter Tuning Regularization, Dropout etc.

email_classification_nlp

corpus_wordnet_textclassification

nltk_tokenizer_stem_stopwrd_speechtag

restaurent_review_sentiment