DBSCAN Clustering Algorithm #1036
Conversation
|
I'm not sure why but some test unrelated to this PR don't make it (ObjectDetector..) |
|
@asvsfs - This is amazing! Wow super cool. This has been on my list for a while now and am so glad you picked this up. I won't have time to review for awhile in depth, but I will let you know when I am able to test and give feedback. On a quick first glance, things are looking good! I'm excited to have a look. Thanks for your contribution! |
src/DBSCAN/index.js
Outdated
| }; | ||
| this.lastClusterId = 0; | ||
| this.status = []; | ||
| this.load(dataset).then(callback); |
There was a problem hiding this comment.
Hi! I'm wondering why there are two calls to this.load() on lines 67 and 68. Is it possible that you only wanted to call this once?
If so, I think line 67 does not need to be included. We can just keep line 68.
There was a problem hiding this comment.
@joeyklee Corrected.
| // Step 4: | ||
| // use the fancy d3 to make magic | ||
| function makeChart() { | ||
| const dataset = dbscanModel.dataset; |
There was a problem hiding this comment.
Small suggestion: we can use object destructuring here:
const { dataset } = dbscanModel;
|
NOTE: I will open an issue for a bug I noticed with how our examples.json gets created. I found that the dbscan examples weren't showing for some reason and it is because the |
|
@all-contributors please add @asvsfs for code example |
|
I've put up a pull request to add @asvsfs! 🎉 |
Dear ml5 community,
I'm making a Pull Request(PR). Please see the details below.
A good PR 🌟
→ Step 1: Which branch are you submitting to? 🌲
Development
→ Step 2: Describe your Pull Request 📝
I've added a simple DBSCAN clustering algorithm to the models , it is influenced by jDBSCAN and WIKIPEDIA definition of the algorithm
A great PR 🌟🌟
→ Step 3: Share a Relevant Example 🦄
D3 Example is included in the project.
The best PR 🌟🌟🌟
→ Step 4: Screenshots or Relevant Documentation 🖼
I've tried to Tensorized it as much as i could since i'm completely new to Tensorflow , any help is welcome to improve the code 👍
