We developed a Java Pinyin Input Method. The most important difference between it and current input methods is that it integrates more syntax information, n-gram and dependency parsing.
  • n-gram is the most widely used language model which are integrated in the products of Sougou IME, Google IME and MS IME.
  • dependency parsing can capture the long distance information in the sentence. We expect to bring it into the n-gram.
This website is just a STARTING POINT.

The snapshot of current Java IME can be found here.

The source code can be found here.

The Google Code Project can be found here. (not update since Mar. 2009)

Note: the source code was not updated since Feb. 2010

Yifan Peng,
Jun 9, 2009, 8:37 AM
Yifan Peng,
Feb 27, 2010, 1:55 AM