librahu / herec Goto Github PK
View Code? Open in Web Editor NEWSource code for TKDE 2018 "Heterogeneous information network embedding for recommendation"
Source code for TKDE 2018 "Heterogeneous information network embedding for recommendation"
Hello, the article mentions the Douban dataset and Douban book dataset. Can you provide these datasets?
Hi author:
I feel confused about the dataset you provide. can you tell me the meaning of the data set? Thank you very much!
What do you mean by data in files bca, BCI , ub,uco and uu? What should I do if I want to change my data?
你好,请问在rating prediction这个地方,x和y是如何获得的呢?我的理解e是融合后得到的向量,r可能是随机初始化需要学习的一个向量表示吗?请问x、y和r他们的含义是什么呢?
你好作者大大,我这边有个疑问哈,我通过你的源码,跑了下cut_data,生成的bcab文件一共三列,如下:
1 4 1
1 2 3
1 23 5
4 1 1
4 11 1
...
每一行代表b -> ca -> b 这个元路径吧,如果我理解的没问题的话,第一列b和第二列ca是不能有重复的编码的?
不然这样的话在随机游走的过程中,图就会失去本来的结构,被搞乱了。
还是说deepwalk可以识别出第一列和第二列代表不同类型的节点,可以自动识别呢?
Hi sir,,thanks for your code.
I have a question is that in the article after methapath produced , there is type constraint and filtering on methapaths(part 4.1.2 Type Constraint and Filtering), but in embedding code, i cannot find this part?
do you implement this part? or you write this code for HERecdw as explained in part 5.3 at the end of page 8?
论文中MF计算评分矩阵时提及了基础MF中的item embedding,HIN的item embedding和隐向量embedding三部分,请问,完成训练后,哪个作为item的embedding表示?谢谢。
NOTHING
你好,按照deepwalk的官方指示安装之后运行,显示deepwalk: command not found是什么问题呢? Python版本号是3.7.0, 操作系统Mac OS 10.14.2
Hi sir, thanks for your code. i have a question. I run metapath generation and metapath embedding code on my own dataset. I decide to use simple linear fusion function for my dataset but i only need the fused embedding, i donot need to recommend something. for example i need fused embedding for user10. which part of HERec_sl should i use?
刚刚成为推荐入门者,跑您的程序对一些数据输入(例如 bca.txt bci.txt)
做一个说明呢 (也可以说明引用的数据集及其简写对照)方便大家对照学习。
thx~
HERec训练完成后,计算HIN的embedding,即e(v)的cosine相似度,结果全是0.9999,请问这是为什么?
Hello professor,thanks for your code.
I have a question is that ,when I use DB movie datasets, I change the metapaths ,but when I run HERec ,it has keyError, so, how to config the 'ratedim','itemdim' and 'userdim'? should I change it when I use a bigger dataset?
您好,最近在研读您的论文有几个疑问,通过融合得到的user和item的表示向量是在同一个空间下的吗?为什么不能直接通过计算user和item向量的相似度来做推荐呢?如果我只用元路径约束来随机游走得到序列,而不去做类型约束和过滤就是例如:U1D1U2这种只取U1U2,那我得到的向量可不可以直接直接用向量间的相似度来做推荐?
Hello professor,
I have a problem running HERec_pl on the path, where should I modify the code?
不是很了解工程文件deepwalk如何被导入到herec里面,希望告知。
Hello, sir.
I have run your code on my dataset. My metapath is UPDPU. (user- project-discipline). And I want to recommend co-workers to the applicant of project which belonging to the discipline. But I just got the matrix which row is user, column is project and data is the rating. The rating is meaningless.
I want get the target matrix which row is user, column is user and data is the similarities between users. How can i adjust the input to get that matrix?
您好!
我想问一下您的HERec_pl文件里是怎么体现personalized preference的呢?我感觉cal_u和cal_v的代码与HERec_sl文件里是相同的
您好,最近在研读您的论文,对您的代码想请教一下。对于embedding的方式,您提供的代码只是采用了deepwalk的方式表示节点,并没有提供论文中HERec的方法是吗?
Hellow,sir.
In the process of running the program,I encountered the following questions,can you give me an answer?
train_rate: 0.8
ratedim: 10 userdim: 30 itemdim: 10
max_steps: 100
delta: 0.02 beta_e: 0.1 beta_h: 0.1 beta_p: 2 beta_w: 0.1 beta_b 0.1 reg_u 1.0 reg_v 1.0
Traceback (most recent call last):
File "E:/ler/HERec-master/code/HERec_spl.py", line 261, in
HNERec(unum, inum, ratedim, userdim, itemdim, user_metapaths, item_metapaths, trainfile, testfile, steps, delta, beta_e, beta_h, beta_p, beta_w, beta_b, reg_u, reg_v)
File "E:/ler/HERec-master/code/HERec_spl.py", line 29, in init
self.X, self.user_metapathdims = self.load_embedding(user_metapaths, unum)
File "E:/ler/HERec-master/code/HERec_spl.py", line 53, in load_embedding
with open(sourcefile) as infile:
FileNotFoundError: [Errno 2] No such file or directory: '../data/embedding/ubu_0.8.embedding'
Process finished with exit code 1
Thank you,very much.
你好,请问可否方便提供一下另外两个数据集相对的代码?
Hi, if I don't misunderstand, the data in this repo is Douban (book) in paper? But after I run the code, I get the MSE ~0.88 instead of ~0.55 in paper. Is there anything wrong?
为什么我跑你的代码,结果远远没有论文中说的那么好。基本上只能达到0.8多一些。
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.