Theme Classification Base on Khmer Keywords Search using Naive Bayes Method

The purpose of this project is for thesis.

Environment Requirement

• Python version >= 3.7

• MySQL

• Pycharm CE IDE

Set up libraries and connector

• MySQL connector

pip search mysql-connector | grep --color mysql-connector-python

• In case you need to update sklearn to latest version

pip install -U scikit-learn

Running the demo application

• Running Python as server

cd ROOT_DIRECTORY
python3 server.py

• Running web application. Open the filename demo/index.html with chrome browser

Add more training data

• Insert the link from [https://www.postkhmer.com/](Phnom Penh Post Khmer Website)

Note: Copy the shortlink of the post from the category you want to insert

• Open the stan-alone application standalone-app/BI Narin.zip in order to perform Khmer word segmentation

Note: You can manually insert white-space, hidden space or empty space, to the document but it is not recommended.

• Upzip the file

• Setting up support library from standalone-app/BI Narin/Setup Packages/KWSAddinSetup(x64).msi

Note: Install library that support with our current operating system.

• Run the standalone-app/BI Narin/Setup Packages/KWSAppSetup(x64).msi file

Note: Run the .exe file that support with your current operating system.

• Running crawl

python3 crawler.py

Note: If there are some error, please try to re-run it because we crawl more than 1000 posts and it could some how was block.

• Go to web application to train the inserted data

Add more cleaning character

• Edit file clean_word_list.txt

• Add all the characters