Awesome

<p float="left"><img src="https://img.shields.io/badge/python-v3.7-red"> <img src="https://img.shields.io/badge/tensorflow-v1.14-blue"> <img alt="GitHub last commit" src="https://img.shields.io/github/last-commit/Coder-Yu/QRec"></p> <h2>Introduction</h2>

QRec is a Python framework for recommender systems (Supported by Python 3.7.4 and Tensorflow 1.14+) in which a number of influential and newly state-of-the-art recommendation models are implemented. QRec has a lightweight architecture and provides user-friendly interfaces. It can facilitate model implementation and evaluation. <br> Founder and principal contributor: @Coder-Yu <br> Other contributors: @DouTong @Niki666 @HuXiLiFeng @BigPowerZ @flyxu<br> Supported by: @AIhongzhi (<a href="https://sites.google.com/view/hongzhi-yin/home">A/Prof. Hongzhi Yin</a>, UQ), @mingaoo (<a href="http://www.cse.cqu.edu.cn/info/2096/3497.htm">A/Prof. Min Gao</a>, CQU) <br> <br> We also provide Pytorch implementations of some models in another library. Please click here

<h2>What's New</h2> <p> 31/03/2022 - SimGCL proposed in our SIGIR'22 paper has been added. <br> 12/10/2021 - BUIR proposed in SIGIR'21 paper has been added. <br> 30/07/2021 - We have transplanted QRec from py2 to py3. <br> 07/06/2021 - SEPT proposed in our KDD'21 paper has been added. <br> 16/05/2021 - SGL proposed in SIGIR'21 paper has been added. <br> 16/01/2021 - MHCN proposed in our WWW'21 paper has been added.<br> 22/09/2020 - DiffNet proposed in SIGIR'19 has been added. <br> 19/09/2020 - DHCF proposed in KDD'20 has been added. <br> 29/07/2020 - ESRF proposed in my TKDE paper has been added. <br> 23/07/2020 - LightGCN proposed in SIGIR'20 has been added. <br> 17/09/2019 - NGCF proposed in SIGIR'19 has been added. <br> 13/08/2019 - RSGAN proposed in ICDM'19 has been added.<br> 09/08/2019 - Our paper is accepted as full research paper by ICDM'19. <br> 20/02/2019 - IRGAN proposed in SIGIR'17 has been added. <br> 12/02/2019 - CFGAN proposed in CIKM'18 has been added.<br> </p> <h2>Architecture</h2>

QRec Architecture

<h2>Workflow</h2>

QRec Architecture

<h2>Features</h2> <ul> <li><b>Cross-platform</b>: QRec can be easily deployed and executed in any platforms, including MS Windows, Linux and Mac OS.</li> <li><b>Fast execution</b>: QRec is based on Numpy, Tensorflow and some lightweight structures, which make it run fast.</li> <li><b>Easy configuration</b>: QRec configs recommenders with a configuration file and provides multiple evaluation protocols.</li> <li><b>Easy expansion</b>: QRec provides a set of well-designed recommendation interfaces by which new algorithms can be easily implemented.</li> </ul> <h2>Requirements</h2> <ul> <li>gensim==4.1.2</li> <li>joblib==1.1.0</li> <li>mkl==2022.0.0</li> <li>mkl_service==2.4.0</li> <li>networkx==2.6.2</li> <li>numba==0.53.1</li> <li>numpy==1.20.3</li> <li>scipy==1.6.2</li> <li>tensorflow==1.14.0</li> </ul> <h2>Usage</h2> <p>There are two ways to run the recommendation models in QRec:</p> <ul> <li>1.Configure the xx.conf file in the directory named config. (xx is the name of the model you want to run)</li> <li>2.Run main.py.</li> </ul> <p>Or</p> <ul> <li>Follow the codes in snippet.py.</li> </ul>

For more details, we refer you to the handbook of QRec.

<h2>Configuration</h2> <h3>Essential Options</h3> <div> <table class="table table-hover table-bordered"> <tr> <th width="12%" scope="col"> Entry</th> <th width="23%" class="conf" scope="col">Example</th> <th width="65%" class="conf" scope="col">Description</th> </tr> <tr> <td>ratings</td> <td>D:/MovieLens/100K.txt</td> <td>Set the file path of the dataset. Format: each row separated by empty, tab or comma symbol. </td> </tr> <tr> <td>social</td> <td>D:/MovieLens/trusts.txt</td> <td>Set the file path of the social dataset. Format: each row separated by empty, tab or comma symbol. </td> </tr> <tr> <td scope="row">ratings.setup</td> <td>-columns 0 1 2</td> <td>-columns: (user, item, rating) columns of rating data are used.<br> </td> </tr> <tr> <td scope="row">social.setup</td> <td>-columns 0 1 2</td> <td>-columns: (trustor, trustee, weight) columns of social data are used.<br> </td> </tr> <tr> <td scope="row">mode.name</td> <td>UserKNN</td> <td>name of the recommendation model. <br> </td> </tr> <tr> <td scope="row">evaluation.setup</td> <td>-testSet ./dataset/test.txt </td> <td>Main option: -testSet, -ap, -cv (choose one of them) <br> -testSet path/to/test/file (need to specify the test set manually)<br> -ap ratio (ap means that the ratings are automatically partitioned into training set and test set, the number is the ratio of the test set. e.g. -ap 0.2)<br> -cv k (-cv means cross validation, k is the number of the fold. e.g. -cv 5)<br> -predict path/to/user list/file (predict for a given list of users without evaluation; need to mannually specify the user list file (each line presents a user)) <br> Secondary option:-b, -p, -cold, -tf, -val (multiple choices) <br> <b>-val ratio </b> (model test would be conducted on the validation set which is generated by randomly sampling the training dataset with the given ratio.)<br> -b thres （binarizing the rating values. Ratings equal or greater than thres will be changed into 1, and ratings lower than thres will be left out. e.g. -b 3.0）<br> -p (if this option is added, the cross validation wll be executed parallelly, otherwise executed one by one) <br> <b>-tf </b> (model training will be conducted on TensorFlow (only applicable and needed for shallow models)) <br> -cold thres (evaluation on cold-start users; users in the training set with rated items more than thres will be removed from the test set) </td> </tr> <tr> <td scope="row">item.ranking</td> <td>off -topN -1 </td> <td>Main option: whether to do item ranking<br> -topN N1,N2,N3...: the length of the recommendation list. *QRec can generate multiple evaluation results for different N at the same time<br> </td> </tr> <tr> <td scope="row">output.setup</td> <td>on -dir ./Results/</td> <td>Main option: whether to output recommendation results<br> -dir path: the directory path of output results. </td> </tr> </table> </div> <h3>Memory-based Options</h3> <div> <table class="table table-hover table-bordered"> <tr> <td scope="row">similarity</td> <td>pcc/cos</td> <td>Set the similarity method to use. Options: PCC, COS;</td> </tr> <tr> <td scope="row">num.neighbors</td> <td>30</td> <td>Set the number of neighbors used for KNN-based algorithms such as UserKNN, ItemKNN. </td> </tr> </table> </div> <h3>Model-based Options</h3> <div> <table class="table table-hover table-bordered"> <tr> <td scope="row">num.factors</td> <td>5/10/20/number</td> <td>Set the number of latent factors</td> </tr> <tr> <td scope="row">num.max.epoch</td> <td>100/200/number</td> <td>Set the maximum number of epoch for iterative recommendation algorithms. </td> </tr> <tr> <td scope="row">learnRate</td> <td>-init 0.01 -max 1</td> <td>-init initial learning rate for iterative recommendation algorithms; <br> -max: maximum learning rate (default 1);<br> </td> </tr> <tr> <td scope="row">reg.lambda</td> <td>-u 0.05 -i 0.05 -b 0.1 -s 0.1</td> <td> -u: user regularizaiton; -i: item regularization; -b: bias regularizaiton; -s: social regularization</td> </tr> </table> </div> <h2>Implement Your Model</h2> <ul> <li>1.Make your new algorithm generalize the proper base class.</li> <li>2.Reimplement some of the following functions as needed.</li> </ul>           - readConfiguration()<br>           - printAlgorConfig()<br>           - initModel()<br>           - trainModel()<br>           - saveModel()<br>           - loadModel()<br>           - predictForRanking()<br>           - predict()<br> <br>

For more details, we refer you to the handbook of QRec.

<h2>Implemented Algorithms</h2> <div> <table class="table table-hover table-bordered"> <tr> <th>Rating prediction</th> <th>Paper</th> </tr> <tr> <td scope="row">SlopeOne</td> <td>Lemire and Maclachlan, Slope One Predictors for Online Rating-Based Collaborative Filtering, SDM'05.<br> </td> </tr> <tr> <td scope="row">PMF</td> <td>Salakhutdinov and Mnih, Probabilistic Matrix Factorization, NIPS'08. </td> </tr> <tr> <td scope="row">SoRec</td> <td>Ma et al., SoRec: Social Recommendation Using Probabilistic Matrix Factorization, SIGIR'08. </td> </tr> <tr> <td scope="row">SVD++</td> <td>Koren, Factorization meets the neighborhood: a multifaceted collaborative filtering model, SIGKDD'08. </td> </tr> <tr> <td scope="row">RSTE</td> <td>Ma et al., Learning to Recommend with Social Trust Ensemble, SIGIR'09. </td> </tr> <tr> <td scope="row">SVD</td> <td>Y. Koren, Collaborative Filtering with Temporal Dynamics, SIGKDD'09. </td> </tr> <tr> <td scope="row">SocialMF</td> <td>Jamali and Ester, A Matrix Factorization Technique with Trust Propagation for Recommendation in Social Networks, RecSys'10. </td> </tr> <tr> <td scope="row">EE</td> <td>Khoshneshin et al., Collaborative Filtering via Euclidean Embedding, RecSys'10. </td> </tr> <tr> <td scope="row">SoReg</td> <td>Ma et al., Recommender systems with social regularization, WSDM'11. </td> </tr> <tr> <td scope="row">LOCABAL</td> <td>Tang, Jiliang, et al. Exploiting local and global social context for recommendation, AAAI'13. </td> </tr> <tr> <td scope="row">SREE</td> <td>Li et al., Social Recommendation Using Euclidean embedding, IJCNN'17. </td> </tr> <tr> <td scope="row">CUNE-MF</td> <td>Zhang et al., Collaborative User Network Embedding for Social Recommender Systems, SDM'17. </td> </table> <br> <table class="table table-hover table-bordered"> <tr> <th>Item Ranking</th> <th>Paper</th> </tr> <tr> <td scope="row">BPR</td> <td>Rendle et al., BPR: Bayesian Personalized Ranking from Implicit Feedback, UAI'09.<br> </td> </tr> <tr> <td scope="row">WRMF</td> <td>Yifan Hu et al.Collaborative Filtering for Implicit Feedback Datasets, KDD'09. </td> </tr> <tr> <td scope="row">SBPR</td> <td>Zhao et al., Leveraing Social Connections to Improve Personalized Ranking for Collaborative Filtering, CIKM'14<br> </td> </tr> <tr> <td scope="row">ExpoMF</td> <td>Liang et al., Modeling User Exposure in Recommendation, WWW''16.<br> </td> </tr> <tr> <td scope="row">CoFactor</td> <td>Liang et al., Factorization Meets the Item Embedding: Regularizing Matrix Factorization with Item Co-occurrence, RecSys'16. </td> </tr> <tr> <td scope="row">TBPR</td> <td>Wang et al. Social Recommendation with Strong and Weak Ties, CIKM'16'. </td> </tr> <tr> <td scope="row">CDAE</td> <td>Wu et al., Collaborative Denoising Auto-Encoders for Top-N Recommender Systems, WSDM'16'.<br> </td> </tr> <tr> <td scope="row">DMF</td> <td>Xue et al., Deep Matrix Factorization Models for Recommender Systems, IJCAI'17'.<br> </td> </tr> <tr> <td scope="row">NeuMF</td> <td>He et al. Neural Collaborative Filtering, WWW'17. </td> </tr> <tr> <td scope="row">CUNE-BPR</td> <td>Zhang et al., Collaborative User Network Embedding for Social Recommender Systems, SDM'17'. </td> </tr> <tr> <td scope="row">IRGAN</td> <td>Wang et al., IRGAN: A Minimax Game for Unifying Generative and Discriminative Information Retrieval Models, SIGIR'17'.<br> </td> </tr> <tr> <td scope="row">SERec</td> <td>Wang et al., Collaborative Filtering with Social Exposure: A Modular Approach to Social Recommendation, AAAI'18'.<br> </td> </tr> <tr> <td scope="row">APR</td> <td>He et al., Adversarial Personalized Ranking for Recommendation, SIGIR'18'.<br> </td> </tr> <tr> <td scope="row">IF-BPR</td> <td>Yu et al. Adaptive Implicit Friends Identification over Heterogeneous Network for Social Recommendation, CIKM'18'. </td> </tr> <tr> <td scope="row">CFGAN</td> <td>Chae et al. CFGAN: A Generic Collaborative Filtering Framework based on Generative Adversarial Networks, CIKM'18. </td> <tr> <td scope="row">NGCF</td> <td>Wang et al. Neural Graph Collaborative Filtering, SIGIR'19'. </td> </tr> <tr> <td scope="row">DiffNet</td> <td>Wu et al. A Neural Influence Diffusion Model for Social Recommendation, SIGIR'19'. </td> </tr> <tr> <td scope="row">RSGAN</td> <td>Yu et al. Generating Reliable Friends via Adversarial Learning to Improve Social Recommendation, ICDM'19'. </td> </tr> <tr> <td scope="row">LightGCN</td> <td>He et al. LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation, SIGIR'20. </td> </tr> <tr> <td scope="row">DHCF</td> <td>Ji et al. Dual Channel Hypergraph Collaborative Filtering, KDD'20. </td> </tr> <tr> <td scope="row">ESRF</td> <td>Yu et al. Enhancing Social Recommendation with Adversarial Graph Convlutional Networks, TKDE'20. </td> </tr> <tr> <td scope="row">MHCN</td> <td>Yu et al. Self-Supervised Multi-Channel Hypergraph Convolutional Network for Social Recommendation, WWW'21. </td> </tr> <tr> <td scope="row">SGL</td> <td>Wu et al. Self-supervised Graph Learning for Recommendation, SIGIR'21. </td> </tr> <tr> <td scope="row">SEPT</td> <td>Yu et al. Socially-Aware Self-supervised Tri-Training for Recommendation, KDD'21. </td> </tr> <tr> <td scope="row">BUIR</td> <td>Lee et al. Bootstrapping User and Item Representations for One-Class Collaborative Filtering, SIGIR'21. </td> </tr> <tr> <td scope="row">SimGCL</td> <td>Yu et al. Are Graph Augmentations Necessary? Simple Graph Contrastive Learning for Recommendation, SIGIR'22. </td> </tr> </table> </div> <h2>Related Datasets</h2> <div> <table class="table table-hover table-bordered"> <tr> <th rowspan="2" scope="col">Data Set</th> <th colspan="5" scope="col" class="text-center">Basic Meta</th> <th colspan="3" scope="col" class="text-center">User Context</th> </tr> <tr> <th class="text-center">Users</th> <th class="text-center">Items</th> <th colspan="2" class="text-center">Ratings (Scale)</th> <th class="text-center">Density</th> <th class="text-center">Users</th> <th colspan="2" class="text-center">Links (Type)</th> </tr> <tr> <td><a href="https://pan.baidu.com/s/1qY7Ek0W" target="_blank"><b>Ciao</b></a> [1]</td> <td>7,375</td> <td>105,114</td> <td width="6%">284,086</td> <td width="10%">[1, 5]</td> <td>0.0365%</td> <td width="4%">7,375</td> <td width="5%">111,781</td> <td>Trust</td> </tr> <tr> <td><a href="http://www.trustlet.org/downloaded_epinions.html" target="_blank"><b>Epinions</b></a> [2]</td> <td>40,163</td> <td>139,738</td> <td width="6%">664,824</td> <td width="10%">[1, 5]</td> <td>0.0118%</td> <td width="4%">49,289</td> <td width="5%">487,183</td> <td>Trust</td> </tr> <tr> <td><a href="https://pan.baidu.com/s/1hrJP6rq" target="_blank"><b>Douban</b></a> [3]</td> <td>2,848</td> <td>39,586</td> <td width="6%">894,887</td> <td width="10%">[1, 5]</td> <td>0.794%</td> <td width="4%">2,848</td> <td width="5%">35,770</td> <td>Trust</td> </tr> <tr> <td><a href="http://files.grouplens.org/datasets/hetrec2011/hetrec2011-lastfm-2k.zip" target="_blank"><b>LastFM</b></a> [4]</td> <td>1,892</td> <td>17,632</td> <td width="6%">92,834</td> <td width="10%">implicit</td> <td>0.27%</td> <td width="4%">1,892</td> <td width="5%">25,434</td> <td>Trust</td> </tr> <tr> <td><a href="https://www.dropbox.com/sh/h97ymblxt80txq5/AABfSLXcTu0Beib4r8P5I5sNa?dl=0" target="_blank"><b>Yelp</b></a> [5]</td> <td>19,539</td> <td>21,266</td> <td width="6%">450,884</td> <td width="10%">implicit</td> <td>0.11%</td> <td width="4%">19,539</td> <td width="5%">864,157</td> <td>Trust</td> </tr> <tr> <td><a href="https://www.dropbox.com/sh/20l0xdjuw0b3lo8/AABBZbRg9hHiN42EHqBSvLpta?dl=0" target="_blank"><b>Amazon-Book</b></a> [6]</td> <td>52,463</td> <td>91,599</td> <td width="6%">2,984,108</td> <td width="10%">implicit</td> <td>0.11%</td> <td width="4%">-</td> <td width="5%">-</td> <td>-</td> </tr> </table> </div> <h3>Reference</h3> <p>[1]. Tang, J., Gao, H., Liu, H.: mtrust:discerning multi-faceted trust in a connected world. In: International Conference on Web Search and Web Data Mining, WSDM 2012, Seattle, Wa, Usa, February. pp. 93–102 (2012)</p> <p>[2]. Massa, P., Avesani, P.: Trust-aware recommender systems. In: Proceedings of the 2007 ACM conference on Recommender systems. pp. 17–24. ACM (2007) </p> <p>[3]. G. Zhao, X. Qian, and X. Xie, “User-service rating prediction by exploring social users’ rating behaviors,” IEEE Transactions on Multimedia, vol. 18, no. 3, pp. 496–506, 2016.</p> <p>[4]. Iván Cantador, Peter Brusilovsky, and Tsvi Kuflik. 2011. 2nd Workshop on Information Heterogeneity and Fusion in Recom- mender Systems (HetRec 2011). In Proceedings of the 5th ACM conference on Recommender systems (RecSys 2011). ACM, New York, NY, USA</p> <p>[5]. Yu et al. Self-Supervised Multi-Channel Hypergraph Convolutional Network for Social Recommendation, WWW'21.</p> <p>[6]. He et al. LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation, SIGIR'20.</p> <h2>Acknowledgment</h2> <p>This project is supported by the Responsible Big Data Intelligence Lab (RBDI) at the school of ITEE, University of Queensland, and Chongqing University.</p>

If our project is helpful to you, please cite one of these papers.<br> <br> @inproceedings{yu2021socially,<br> title={Socially-aware self-supervised tri-training for recommendation},<br> author={Yu, Junliang and Yin, Hongzhi and Gao, Min and Xia, Xin and Zhang, Xiangliang and Viet Hung, Nguyen Quoc},<br> booktitle={Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining},<br> pages={2084--2092},<br> year={2021}<br> } <br> <br> @inproceedings{yu2021self,<br> title={Self-Supervised Multi-Channel Hypergraph Convolutional Network for Social Recommendation},<br> author={Yu, Junliang and Yin, Hongzhi and Li, Jundong and Wang, Qinyong and Hung, Nguyen Quoc Viet and Zhang, Xiangliang},<br> booktitle={Proceedings of the Web Conference 2021},<br> pages={413--424},<br> year={2021}<br> }