search

capstone496 / SpeechSentiments

1 stars 1 forks source link

Sentiment Analysis Algorithm #3

Open rightnknow opened 6 years ago

rightnknow commented 6 years ago

The Algorithm will target on balancing between accuracy and efficiency.

Initial thoughts:

  1. split datasets into two section: 1. raw audio data 2. text data
  2. Train the model, tune the parameter,
  3. weight the results taken from two model(can be more if we have different model for each type of data)
KatJHuang commented 6 years ago

Sections 6 & 8 of this paper (https://www.sciencedirect.com/science/article/pii/S0925231215011297) provide a simple way of fusing features for multimodal content.

KatJHuang commented 5 years ago

Issue #5 explores the option of CNN on audio data.