# Amazon-Apps Datasets

## Dataset Description
The Amazon-Apps dataset is a shopping network. It includes information about items, users and reviews. Nodes represent items and users. Text on a user node is the `reviwer` and on a item node is `item`. Edges represent relationships between items and users. Text on an edge is a user's review of a item including following information: `Reviewer [reviewerName] left a review on [reviewTime], giving the product [rating] stars. In his/her review, he/she wrote: [reviewText]. His/Her summary was [summary].`. 


## Graph Machine Learning Tasks

### Link Prediction
Link prediction in the Amazon-Apps dataset involves predicting potential connections between users and items.  The goal is to predict whether a user will purchase a item.

### Node Classification
Node classification tasks in the Amazon-Apps dataset include predicting the items' category. 


## Dataset Source
http://jmcauley.ucsd.edu/data/amazon/links.html