# Ada_SR

**Repository Path**: meizi_camellia/ada_sr

## Basic Information

- **Project Name**: Ada_SR
- **Description**: No description available
- **Primary Language**: Unknown
- **License**: ISC
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2024-11-23
- **Last Updated**: 2024-11-23

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README





# Adaptive User Multi-level and Multi-Interest Preferences for Sequential Recommendation



## 1. Introduction

We proposes an adaptive user multi-level and multi-interest preferences for sequential recommendation (AdaSR) that decouples the learning of users' local sequential and global collaborative multi-interest preferences. Specifically, AdaSR consists of a preference encoding module, a user multi-interest decoupling and merging module, and a contrastive learning assistance task. We first learn user local sequential and global collaborative preferences separately by preference encoding module, which can effectively encode preference features in user-interacted sequences and similar users. To obtain the proper number of multi-interests, we propose a user multi-interest decoupling and merging module, which decouples the user's local and global preferences into multi-interest preferences. It removes redundant interests among them and adaptively learns the number of multi-interest preferences for each user.
![model](fig2.jpg)

## 2. Running environment

We develop our codes in the following environment:

1. Ubuntu OS
2. Python >= 3.7.9
3. torch 1.4.0+
4. Nvidia GPU with cuda 10.1+

## 3. Datasets
Three popular public datasets for recommendation are used in our research:
1. Musical Instruments
2. Toy and Games

| Dataset      | # User | # Item | # Interaction | 
| ------------ | ------ | ------ | ------------- | 
| Musical Instruments      |60,739  | 56,301 | 133,189       |
| Toy and Games        | 313,557| 241,657 | 784,844       |

## 4. How to run the codes

python train.py


## 5. Citation

If you find this work helpful to your research, please kindly consider citing our paper.

```
XXX
```