This repository contains two PyQt5-based tools for Chinese Named Entity Recognition (NER) and tag frequency analysis.

• Select a training dataset and train a CRF model
• Select raw input text and predict NER tags
  • Supports Windows and macOS/Linux
    • Displays prediction results in the output box

• Input text in word/tag format (e.g., 香港/ns)
• Enter a target tag (e.g., nr) to analyze
• Visualizes the top 20 words for that tag
• Supports Chinese fonts via .ttf file

Install required packages:

pip install pyqt5 matplotlib

Make sure the crf/ folder contains:

• crf_learn.exe / crf_learn
• crf_test.exe / crf_test
• template file for training

Optional: Chinese font MingLiU.ttf for plotting.

Run the GUI:

python crf_ner_gui.py

Steps:

1. Select your training dataset (.txt)
2. Train the CRF model
3. Select raw input text
4. Predict NER tags

Run the GUI:

python tag_frequency_gui.py

Steps:

1. Paste your word/tag text in the text area
2. Enter a target tag (e.g., nr)
3. Click Plot Frequency to visualize the top 20 words

Formatted input for CRF NER GUI:

直	O
至	O
十	O
多	O
年	O
前	O
﹐	O
梁	S-nr
購	O
入	O
兇	O
案	O
現	O
場	O
深	B-nr
水	I-nr
	E-nr
北	B-nr
河	I-nr
街	E-nr
十	O
三	O
號	O
二	O
樓	O

Sample word/tag text for Tag Frequency Statistics:

梁/S-nr 深/B-nr 水/I-nr /E-nr 北/B-nr 河/I-nr 街/E-nr

Prediction output:

梁/S-nr 深/B-nr 水/I-nr /E-nr 北/B-nr 河/I-nr 街/E-nr

This project is released under the MIT License.

Created by [Your Name] – email: youremail@example.com

HKUST official website