一个基于java的语音标签读取和写入部分门兼容的XML.zip

I. The Assisted Tagger – an Introduction

We would like to compare annotations made with pre-tagged words (assisted)

against annotations without any pre-tagged words (unassisted). To do that, we

have designed an tagger interface that can be used in assisted or unassisted

mode.

Here’s a brief user’s guide to the tagger interface. It has two modes of operation

– one assisted, one unassisted. Most of the functionality is the same in either

mode, but there are a few important differences. First, a brief overview of both

modes:

mode, a list of parts of speech takes up the right pane, and all tags are marked

as unknown. The human tagger will change the unknown tags to the correct

tags.

In the assisted mode, the words are pre-tagged with a bigram tagger trained on

the Wall Street Journal corpus. The right pain contains statistical information

from the training corpus. The human tagger can either leave the tag selected by

the bigram tagger or can change the tag to the correct tag.

For both modes, humans only need to edit part-of-speech tags for words in the

report. All other elements –punctuation marks, dashes, parentheses, and other

non-alphanumerics – will be pre-tagged and not editable.

Figures 1 and 2) and select the correct tag from the alphabetically

organized drop-down list

� type correct tag in current word combo box in the Current Word Pane (A

in Figures 1 and 2), which will jump to the tag you are looking for in the

list

Unassisted Mode:

� Select the correct tag from the Part of Speech list on the right pane (B in

Figure 2)

Assisted Mode:

� Select the correct tag from any list on the right pane (B in Figure 1)