docs: update README (#66)

* Update README.md
* remove convert2plain_body.py
* convert2plain_title.py

README.md

@@ -15,6 +15,33 @@ A GitHub Action to suggest related or similar issues, documents, and links. Base
 
 <!-- END doctoc generated TOC please keep comment here to allow auto update -->
 
+## Models
+
+Our GitHub Action is **actions-suggest-related-links**,  which suggests related or similar issues, documents, and links. 
+This action mainly consists of 5 parts: Data Collection, Preprocessing, Train model, Find similar issues, and Suggest Issues.
+
+![workflow](https://user-images.githubusercontent.com/39023477/93440029-37d64580-f90a-11ea-8e7b-625992aa0a58.jpg)
+
+### Data Collection
+
+All issues of the repository are collected in JSON format based on the `GitHub API`. Issues include the title, body, and comments. Training Data is regularly collected using the scheduling function and output as an artifact and saved as a cache.
+
+### Preprocessing
+
+We convert markdown to text based on `unifiedjs`. At this time, symbols that are not alphabetic characters are deleted.
+
+- https://github.com/peaceiris/actions-suggest-related-links/blob/main/actions/src/preprocess.ts
+
+### Train Model
+
+When the new issue is updated, the model is trained based on `fasttext`. In accordance with its name, fasttest has the advantage of very short inference times. We think training at GitHub Actions won't be an issue of execution time. In fact, in [actions-gh-pages](https://github.com/peaceiris/actions-gh-pages) repository, the training execution time was ???.
+
+### Find Similar Issues
+
+Calculate word vectors of training data and word vectors of posted data in fasttext. The cosine similarity is used to determine which word vectors of training data is close to the word vectors of the posted data. The higher the cosine similarity, the more similar the sentence.
+
+### Suggest Issues
+
 
 
 ## Maintainers