AutoTaggingStackOverFlowData

• MultiClass, Multi Label Classification problem
• Dataset Contains StackOverFlow Data with Question Title, question and the labels
• Steps:
  • Clean the Data
  • Combine Title and Question into a single column and use Tfidf Vectoriser for creating training set
  • Use Sklearn preprocessing MultiLabel Binariser on the target variable to make it into a Classification problem
  • Build model 1 with Logistic Regression and SkLearn Multiclass OnevsRestClassifier
  • Use F1 Score metric for evaluation
  • Build model with deep learning
    • Use Keras Tokenizer and Pad Sequence to create input sequence
    • Using Keras models Sequential, Load Model
    • keras layers Embedding, Dropout, Conv1d, GlobalMaxPool1d, Dense
  • Create a inference Function to test for new data