Abstract:Thanks to the great interest posed by researchers and companies, recommendation systems became a cornerstone of machine learning applications. However, concerns have arisen recently about the need for reproducibility, making it challenging to identify suitable pipelines. Several frameworks have been proposed to improve reproducibility, covering the entire process from data reading to performance evaluation. Despite this effort, these solutions often overlook the role of data management, do not promote interoperability, and neglect data analysis despite its well-known impact on recommender performance. To address these gaps, we propose DataRec, which facilitates using and manipulating recommendation datasets. DataRec supports reading and writing in various formats, offers filtering and splitting techniques, and enables data distribution analysis using well-known metrics. It encourages a unified approach to data manipulation by allowing data export in formats compatible with several recommendation frameworks.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the standardization and reproducibility issues in data processing and analysis in the field of recommendation systems. Specifically, although many frameworks have been proposed to improve the reproducibility and evaluation process of recommendation systems, these frameworks often overlook the following key aspects: 1. **Data Management**: Existing frameworks do not comprehensively handle data reading, writing, filtering, splitting, etc. 2. **Interoperability**: There is a lack of compatibility between different frameworks, making it difficult for researchers to integrate and reuse different tools and techniques. 3. **Data Analysis**: Many frameworks ignore the analysis of the characteristics of recommendation datasets, and these characteristics have a significant impact on the performance of recommendation systems. To solve these problems, the author proposes a new framework named DataRec. The main goal of DataRec is to provide a standardized data processing and analysis platform that supports multiple data formats and processing techniques and can be seamlessly integrated with existing recommendation system frameworks. In this way, DataRec aims to promote the standardization and interoperability of the data management and analysis processes in recommendation system research. ### Specific Problems and Solutions - **Data Management**: - DataRec supports multiple commonly - used data formats (such as CSV, TSV, JSON, etc.) and provides rich data reading and writing functions. - It also implements common data filtering and splitting strategies, such as k - Core filtering, time - splitting, etc. - **Interoperability**: - DataRec is designed as a Python module and can be easily integrated into other workflows without complex configuration files. - It can export the processed datasets, making them compatible with multiple popular recommendation system frameworks. - **Data Analysis**: - DataRec introduces the analysis of the characteristics of recommendation datasets, including the distribution of users and items, the cold - start problem, etc., and uses metrics such as the Gini index for measurement. - These analysis tools can help researchers better understand the characteristics of the datasets, so as to choose more appropriate experimental settings. Through the above measures, DataRec aims to fill the gaps in data processing and analysis in existing frameworks and promote the standardization and transparency in the field of recommendation systems.

DataRec: A Framework for Standardizing Recommendation Data Processing and Analysis

A Novel Framework To Process The Quantity And Quality Of User Behavior Data In Recommender Systems

FairRec: Fairness Testing for Deep Recommender Systems

A Reproducible Analysis of Sequential Recommender Systems

RecBole: Towards a Unified, Comprehensive and Efficient Framework for Recommendation Algorithms

Recommendation Systems in Libraries: an Application with Heterogeneous Data Sources

The TagRec Framework as a Toolkit for the Development of Tag-Based Recommender Systems

RecRec: Algorithmic Recourse for Recommender Systems

BETA-Rec: Build, Evaluate and Tune Automated Recommender Systems.

SocRecM: A Scalable Social Recommender Engine for Online Marketplaces

RePlay: a Recommendation Framework for Experimentation and Production Use

DeepRec: An Open-source Toolkit for Deep Learning based Recommendation

NineRec: A Benchmark Dataset Suite for Evaluating Transferable Recommendation

pyRecLab: A Software Library for Quick Prototyping of Recommender Systems

Requirements Engineering for General Recommender Systems

Data science for next-generation recommender systems

Behind Recommender Systems: the Geography of the ACM RecSys Community

recommenderlab: An R Framework for Developing and Testing Recommendation Algorithms

VizRec: A framework for secure data exploration via visual representation

RecFlow: An Industrial Full Flow Recommendation Dataset

A Decision Tree to Shepherd Scientists through Data Retrievability