memgraph
Add Hits and Personalized PageRank module
Description
Please briefly explain the changes you made here.
Pull request type
- [ initial Hits and Personalized PageRank module ] Algorithm/Module
- [ initial Hits and Personalized PageRank module ] Feature
######################################
Reviewer checklist (the reviewer checks this part)
Module/Algorithm
- [ ] Core algorithm/module implementation
- [ ] Query module implementation
- [ ] Unit tests ######################################
Yes, of course. It is not totally finished. I need a little help putting it in Memgraph.
Yes, of course. It is not totally finished. I need a little help putting it in Memgraph.
What do you need help with, and what is not finished? Let us know so we can help you better!
I implemented the algorithms and tested them like others, so it is alright. But I don't know how to make the model use it inside the Memgraph.
I implemented the algorithms and tested them like others, so it is alright. But I don't know how to make the model use it inside the Memgraph.
Did you try following our docs page to enable Memgraph to use module?
Dear Antonio
I developed the HITS algorithm to a certain extent.
In the case of PageRank, I need your advice. The PageRank that already exists in MAGE is a simple one. It does not consider dangle nodes, personalization, Nstart, Edge weight, or dangling weight (out edges ).
The PageRank I implemented at the moment considers dangle nodes and personalization, and if I were to develop it, it would have Nstart, Edge weight, and Dangle weight.
MAGE can not have both of them (PageRank, personalized PageRank) . I believe there should be one PageRank that contains all features. What would you like PageRank to be?
However, I am concerned about my algorithm's time performance. I would appreciate it if you could check it out and provide me with feedback.
Dear Antonio
I developed the HITS algorithm to a certain extent.
In the case of PageRank, I need your advice. The PageRank that already exists in MAGE is a simple one. It does not consider dangle nodes, personalization, Nstart, Edge weight, or dangling weight (out edges ).
The PageRank I implemented at the moment considers dangle nodes and personalization, and if I were to develop it, it would have Nstart, Edge weight, and Dangle weight.
MAGE can not have both of them (PageRank, personalized PageRank) . I believe there should be one PageRank that contains all features. What would you like PageRank to be?
However, I am concerned about my algorithm's time performance. I would appreciate it if you could check it out and provide me with feedback.
Hi, no problem, we will check the performance of your algorithm and if it is the better version of PageRank, there will be only one version and it should be the better one. For now please connect it first to memgraph, and then we can help by checking performance and any other issues. Does that sound good to you?
Hi @1rezatbz, is something happening here :) do you need any help connecting the memgraph to the algorithm?
Dear Antonio,
I'm sorry for the delay, but I'm searching for a job and having to attend interviews and assessments.
As I told you, I developed the Hits Algorithm connection, and I think it's ready. The problem I have is with PageRank. My page rank algorithm requires the user's personalization vector. The right way to do that isn't clear to me. To get help from other existing algorithms in Mage, I considered other existing algorithms, but I could not find any clues. I would appreciate it if you could look at my PageRank code and give me some suggestions!