# Build Docker Container
docker build -t terra_app .
# Running Docker Container
docker run -p 8501:8501 terra_app
-
Customer Interactions:
- Customer ID
- Page views
- Time spent on the website
-
Purchase History:
- Customer ID
- Product ID
- Purchase date
-
Product Details:
- Product ID
- Category
- Price
- Ratings
Due lack of provided dataset, we using syntentic generator libraries called sdv to generate n-sample of rows according datasets.
Customer has their own similar pattern about product purchases, this pattern might be generate because not much variety from sytentic data generator.
We choose K-Nearest Neigbours in terms of Content Based Recommendation System due lack of variety dataset and limited scope of products.
We choose average precision, recall, and f1-score at 5-shot sampling recommendation data, with result all of them is 100% due dataset doesnt have enough variety.