This repository offers a Cantonese dictionary specifically designed for enhancing Cantonese word segmentation. It allows the integration with the jieba segmentation library, enabling you to use it as a custom dictionary for Cantonese text processing.

The development of this Cantonese dictionary has been informed by the following resources:

• Most Common Cantonese Words Frequency List
• Words.hk Faiman Analysis
• PyCantonese
• FastText Pretrained Vectors

To get started, download the dictionary file from the repository:

Download the file `wordlist.txt`

You can integrate this dictionary into your jieba processing workflow as follows:

import jieba
jieba.load_userdict('wordlist.txt')

Here are some examples demonstrating the improvements in segmentation when using this custom dictionary compared to the default jieba dictionary:

1. Example Sentence:

   "影相呃like真係一件好煩嘅事嚟㗎，不過呃like都唔一"

   • Custom Dictionary Output: [影相, 呃like, 真係, 一件, 好煩, 嘅, 事, 嚟, 不過, 呃like, 都, ...]
   • Jieba Default Dictionary Output: [影相, 呃, like, 真, 係, 一件, 好, 煩, 嘅, 事, 嚟, 㗎, ...]

2. Example Sentence:

   "真係唔好怪人哋叫香港做投訴之都⋯⋯繼有餐廳"

   • Custom Dictionary Output: [真係, 唔好, 怪人, 哋, 叫, 香港, 做, 投訴, 之, 都, 繼, 有, 餐廳, ...]
   • Jieba Default Dictionary Output: [真, 係, 唔, 好, 怪人, 哋, 叫, 香港, 做, 投訴, 之, 都, ⋯⋯, ...]