My solution used a dbscan clustering approach, like most others. I used ax,ay and inv_r0 as parameters which takes the helix circle into account. Many nested loops were used to get a lot of iterations for dbscan with different parameters, but I finalized each track after it was created by dbscan and then extended, rather than doing any reassigning of hits later. I've done a more detailed writeup here:
\n\nhttps://github.com/robert604/robert604.github.io/blob/master/index.md
\n\nThe code is also available on github.
", "messages": [ { "id": "369968", "postDate": "08/14/2018 02:55:07", "content": "My solution used a dbscan clustering approach, like most others. I used ax,ay and inv_r0 as parameters which takes the helix circle into account. Many nested loops were used to get a lot of iterations for dbscan with different parameters, but I finalized each track after it was created by dbscan and then extended, rather than doing any reassigning of hits later. I've done a more detailed writeup here:
\n\nhttps://github.com/robert604/robert604.github.io/blob/master/index.md
\n\nThe code is also available on github.
", "rawMarkdown": "My solution used a dbscan clustering approach, like most others. I used ax,ay and inv_r0 as parameters which takes the helix circle into account. Many nested loops were used to get a lot of iterations for dbscan with different parameters, but I finalized each track after it was created by dbscan and then extended, rather than doing any reassigning of hits later. I've done a more detailed writeup here:\n\nhttps://github.com/robert604/robert604.github.io/blob/master/index.md\n\nThe code is also available on github.", "votes": null }, { "id": "370028", "postDate": "08/14/2018 05:34:52", "content": "Congrats @Robert and thanks for sharing your solution.
", "rawMarkdown": "Congrats @Robert and thanks for sharing your solution.", "votes": null }, { "id": "370807", "postDate": "08/15/2018 13:37:52", "content": "Thanks @YaGana Sheriff-Hussaini.
", "rawMarkdown": "Thanks @YaGana Sheriff-Hussaini.", "votes": null }, { "id": "370815", "postDate": "08/15/2018 13:43:27", "content": "Robert, it would be more interesting for us that you share the principles of your solution in your post, esp how does it differ from other DBSCAN approaches already shared, than expecting us to deep dive in your code up front. That's why you did not get many reaction to your post.
\n\nCongrats on the result still, well done.
", "rawMarkdown": "Robert, it would be more interesting for us that you share the principles of your solution in your post, esp how does it differ from other DBSCAN approaches already shared, than expecting us to deep dive in your code up front. That's why you did not get many reaction to your post.\n\nCongrats on the result still, well done.", "votes": null }, { "id": "370884", "postDate": "08/15/2018 15:50:10", "content": "@CPMP I did explain my approach in the github pages. Here's the link:\nhttps://robert604.github.io/\nIt's the webpage link to what I had posted originally. If there's something not clear in that explanation then let me know and I'll make some changes to it.
", "rawMarkdown": "CPMP I did explain my approach in the github pages. Here's the link:\nhttps://robert604.github.io/\nIt's the webpage link to what I had posted originally. If there's something not clear in that explanation then let me know and I'll make some changes to it.", "votes": null }, { "id": "370893", "postDate": "08/15/2018 15:59:07", "content": "That's what I say: you expect us to follow a link when there is no incentive to do so. If you had just copy pasted your explanation here then you would have got more feedback.
", "rawMarkdown": "That's what I say: you expect us to follow a link when there is no incentive to do so. If you had just copy pasted your explanation here then you would have got more feedback.", "votes": null } ], "comments": [ { "id": 370028, "author_name": "sheriytm", "author_url": "", "post_date": "08/14/2018 05:34:52", "content": "Congrats @Robert and thanks for sharing your solution.
", "votes": null, "replies": [] }, { "id": 370807, "author_name": "robertkag", "author_url": "", "post_date": "08/15/2018 13:37:52", "content": "Thanks @YaGana Sheriff-Hussaini.
", "votes": null, "replies": [] }, { "id": 370815, "author_name": "cpmpml", "author_url": "", "post_date": "08/15/2018 13:43:27", "content": "Robert, it would be more interesting for us that you share the principles of your solution in your post, esp how does it differ from other DBSCAN approaches already shared, than expecting us to deep dive in your code up front. That's why you did not get many reaction to your post.
\n\nCongrats on the result still, well done.
", "votes": null, "replies": [ { "id": 370884, "author_name": "robertkag", "author_url": "", "post_date": "08/15/2018 15:50:10", "content": "@CPMP I did explain my approach in the github pages. Here's the link:\nhttps://robert604.github.io/\nIt's the webpage link to what I had posted originally. If there's something not clear in that explanation then let me know and I'll make some changes to it.
", "votes": null, "replies": [] }, { "id": 370893, "author_name": "cpmpml", "author_url": "", "post_date": "08/15/2018 15:59:07", "content": "That's what I say: you expect us to follow a link when there is no incentive to do so. If you had just copy pasted your explanation here then you would have got more feedback.
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "369968": "My solution used a dbscan clustering approach, like most others. I used ax,ay and inv_r0 as parameters which takes the helix circle into account. Many nested loops were used to get a lot of iterations for dbscan with different parameters, but I finalized each track after it was created by dbscan and then extended, rather than doing any reassigning of hits later. I've done a more detailed writeup here:\n\nhttps://github.com/robert604/robert604.github.io/blob/master/index.md\n\nThe code is also available on github.", "370028": "Congrats @Robert and thanks for sharing your solution.", "370807": "Thanks @YaGana Sheriff-Hussaini.", "370815": "Robert, it would be more interesting for us that you share the principles of your solution in your post, esp how does it differ from other DBSCAN approaches already shared, than expecting us to deep dive in your code up front. That's why you did not get many reaction to your post.\n\nCongrats on the result still, well done.", "370884": "CPMP I did explain my approach in the github pages. Here's the link:\nhttps://robert604.github.io/\nIt's the webpage link to what I had posted originally. If there's something not clear in that explanation then let me know and I'll make some changes to it.", "370893": "That's what I say: you expect us to follow a link when there is no incentive to do so. If you had just copy pasted your explanation here then you would have got more feedback." }, "source": "meta" }