Tasks :

• [ ] design a flow first.
• [ ] Python to write an engine
• [ ] upload to GCP

Questions(Google):

1. recommendation strategy (Collaborative Filtering) - refer to eComm solution (Similar user/Similar item ). == Similar item

users preferences. refer to https://towardsdatascience.com/3-approaches-to-build-a-recommendation-system-ce6a7a404576

1. Architect flow (quick win) . == MySQL+ Python

mySQL : https://blogs.oracle.com/mysql/post/building-your-first-machine-learning-model-with-mysql-heatwave-ml

1. frequency of data engine. (training with new data/ recommendation). == Leave it to next level.

   1.training set 2.test set 3. Evaluate

Data table :

1. Words

2.Relationship

 Source : official / **Customized**

 Connection Type : 1. main - main   2.main - Sub (Simple similar/Hard similar/Opposite)

 Adopted : Check/Uncheck (like it or not)

Memo :