Noisy Channel Score

I. Introduction

Noisy Channel Model is used by combining above scores for the ranking. More investigation and tests are planned to be performed when more resources are available.

• NoisyChannelScore.java
  => NC score = language Model score x error Model score
  where:
  • language model score: word frequency score
  • error model score: orthographic score

II. Results

Tests on non-word on the training set with different ranking mode for the function mode of 1-to-1 and Split

From the results:

• The noisy channel score is better than word frequency score, but worse than Orthographic score.