Skip to content

Latest commit

 

History

History
18 lines (13 loc) · 806 Bytes

File metadata and controls

18 lines (13 loc) · 806 Bytes

ErrorTextClassification

Use Word2vec and Neural Networks to classify the error feedback of the passenger seat screen on the plane. First use Word2vec get the vector for each word, then feed vector of each error info sentence into Nerual Nets.

Word2vec Trained Result

  • Word Vector Plot Word
  • Class Vector Plot Class

Error Text Training Methods

nltk.NaiveBayesClassifier Acc ~ 48%

Softmax Regression Acc ~ 77%

Nerual Networks Acc ~ 97%

Reference