Building a Recommender System with Amazon SageMaker Factorization Machines and BlazingText
Notebook: fm_amazon_recommender.ipynb
- Background (Factorization Machines)
- Setup
- Spin up SageMaker hosted notebook instance in console
- Add SageMaker IAM policy to this SageMaker notebook to allow S3 read/write access
- Create new S3 bucket (first cell)
- Import necessary libraries (second cell)
- Dataset
- Overview
- Source: Amazon Reviews Public Dataset
- Features
- Data preprocessing
- Training
- Create SageMaker estimator
- Launch training job
- Host
- Deploy endpoint to perform inference