naoigcat/khcoder repository overview

⁠Docker KH Coder

Docker Image for KH Coder⁠.

⁠Dependencies

• KH Coder 2.00f
• Ubuntu 18.04
• Perl 5.26.1
• R 3.4.4
• MeCab 0.996
• MeCab IPADic 2.7.0-20070801

⁠Installation

docker pull naoigcat/khcoder

⁠Usage

docker run --rm --detach --publish 5900:5900 --volume ${PWD}:/root/Desktop/work naoigcat/khcoder
open vnc://localhost:5900

• Password for VNC: secret

⁠Tutorial⁠

1. Click 'Menu' > 'Project' > 'New'
2. Click 'Browse'
3. Open /root/Desktop/work/tutorial_jp/kokoro.xls
4. Confirm that テキスト is selected for the column to be analyzed
5. Confirm that 日本語, MeCab is selected for the language
6. Click 'OK'
7. Click 'Menu' > 'PRe-Processing' > 'Select Words to Analyze'
8. Enter 一人 and 二人 to 'Pick up following strings as words:'
   • Since Japanese cannot be entered directly on the KH Coder screen, enter it in Leafpad and then copy and paste.
   • It is possible to force the extraction of words that are not extracted as a single word even though they are important words.
   • It is also effective when the division is too fine, such that 一人 is divided into 一 and 人.
   • It is useful for identifying too fine divisions executing 'Menu' > 'PRe-Processing' > 'Word Clusters'.
9. Click 'OK'
10. Click 'Menu' > 'PRe-Processing' > 'Run Pre-Processing'
11. Click 'OK'

⁠Note

• Since DBD::CSV was not installed the first time, install DBD::CSV is executed twice.
• Since the old version of the R packages could not be installed from official repository, it is installed using the remote package.
• Since the name of the Desktop directory changes depending on the language, it is fixed to English by executing LANG=C xdg-user-dirs-update --force.
• Since Japanese could not be entered directly into KH Coder, it is assumed to be input to Leafpad and copy and paste to KH Coder.