{ "id": 55902, "title": "Conventional Algorithms", "url": "/competitions/trackml-particle-identification/discussion/55902", "author_name": "", "post_date": "2018-05-03T00:45:45.972962200Z", "votes": null, "comment_count": 6, "views": 0, "content": "

The awards tab mentions: “...with respect to the traditional particle physics combinatorial approach...”. As this appears to be crucial to final ranking, can you point to a (hopefully brief) turtorial or description of the combinatorial approach?

", "messages": [ { "id": "322453", "postDate": "05/03/2018 00:45:45", "content": "

The awards tab mentions: “...with respect to the traditional particle physics combinatorial approach...”. As this appears to be crucial to final ranking, can you point to a (hopefully brief) turtorial or description of the combinatorial approach?

", "rawMarkdown": "The awards tab mentions: “...with respect to the traditional particle physics combinatorial approach...”. As this appears to be crucial to final ranking, can you point to a (hopefully brief) turtorial or description of the combinatorial approach?", "votes": null }, { "id": "322460", "postDate": "05/03/2018 01:56:03", "content": "

Hey pete, check out page 20 of this talk. I found it via this Kaggle discussion, posted by David Rousseau, one of the challenge creators.

", "rawMarkdown": "Hey pete, check out [page 20 of this talk][1]. I found it via [this Kaggle discussion][2], posted by David Rousseau, one of the challenge creators.\n\n\n [1]: https://drive.google.com/file/d/1tW0tSJAgPJlQOwoSQPG7mtzDd-08F4z1/view\n [2]: https://www.kaggle.com/c/trackml-particle-identification/discussion/55726#321441", "votes": null }, { "id": "322477", "postDate": "05/03/2018 03:06:17", "content": "

I read that the organisers will release basline code for hough transform and dbscan later. Let me know if i am wrong.

", "rawMarkdown": "I read that the organisers will release basline code for hough transform and dbscan later. Let me know if i am wrong.", "votes": null }, { "id": "322533", "postDate": "05/03/2018 06:24:29", "content": "

The applied method in the two main experiment at the LHC is conceptional very simple:\n- build track seeds from points of three\n- run a (combinatorial) Kalman Filter to complete your trajectories

\n\n

Quite lengthy, but here is some background information:

\n\n

Cornelissen, T. et al., Concepts, Design and Implementation of the ATLAS New Tracking (NEWT), ATL-SOFT-PUB-2007-007, https://cds.cern.ch/record/1020106

\n\n

Adam, W. et al., Track reconstruction in the CMS tracker, CERN-CMS-NOTE-2006-041, CMS-NOTE-2006-041

", "rawMarkdown": "The applied method in the two main experiment at the LHC is conceptional very simple:\n- build track seeds from points of three\n- run a (combinatorial) Kalman Filter to complete your trajectories\n\nQuite lengthy, but here is some background information:\n\n Cornelissen, T. et al., *Concepts, Design and Implementation of the ATLAS New Tracking (NEWT)*, ATL-SOFT-PUB-2007-007, https://cds.cern.ch/record/1020106\n\nAdam, W. et al., *Track reconstruction in the CMS tracker*, CERN-CMS-NOTE-2006-041, CMS-NOTE-2006-041", "votes": null }, { "id": "322545", "postDate": "05/03/2018 06:59:54", "content": "

Thanks - very helpful.

", "rawMarkdown": "Thanks - very helpful.", "votes": null }, { "id": "322546", "postDate": "05/03/2018 07:02:02", "content": "

OK that makes sense. I guess you take every “sensible” three point combinations (that could possible be in a track) and run the Kalman filter on it.

", "rawMarkdown": "OK that makes sense. I guess you take every “sensible” three point combinations (that could possible be in a track) and run the Kalman filter on it.", "votes": null }, { "id": "322548", "postDate": "05/03/2018 07:04:59", "content": "

Indeed - below 3 hits there's simply not enough degrees of freedom to constrain the trajectory.

", "rawMarkdown": "Indeed - below 3 hits there's simply not enough degrees of freedom to constrain the trajectory.", "votes": null } ], "comments": [ { "id": 322460, "author_name": "jbonatt", "author_url": "", "post_date": "05/03/2018 01:56:03", "content": "

Hey pete, check out page 20 of this talk. I found it via this Kaggle discussion, posted by David Rousseau, one of the challenge creators.

", "votes": null, "replies": [ { "id": 322545, "author_name": "petewills", "author_url": "", "post_date": "05/03/2018 06:59:54", "content": "

Thanks - very helpful.

", "votes": null, "replies": [] } ] }, { "id": 322477, "author_name": "hengck23", "author_url": "", "post_date": "05/03/2018 03:06:17", "content": "

I read that the organisers will release basline code for hough transform and dbscan later. Let me know if i am wrong.

", "votes": null, "replies": [] }, { "id": 322533, "author_name": "asalzburger", "author_url": "", "post_date": "05/03/2018 06:24:29", "content": "

The applied method in the two main experiment at the LHC is conceptional very simple:\n- build track seeds from points of three\n- run a (combinatorial) Kalman Filter to complete your trajectories

\n\n

Quite lengthy, but here is some background information:

\n\n

Cornelissen, T. et al., Concepts, Design and Implementation of the ATLAS New Tracking (NEWT), ATL-SOFT-PUB-2007-007, https://cds.cern.ch/record/1020106

\n\n

Adam, W. et al., Track reconstruction in the CMS tracker, CERN-CMS-NOTE-2006-041, CMS-NOTE-2006-041

", "votes": null, "replies": [ { "id": 322546, "author_name": "petewills", "author_url": "", "post_date": "05/03/2018 07:02:02", "content": "

OK that makes sense. I guess you take every “sensible” three point combinations (that could possible be in a track) and run the Kalman filter on it.

", "votes": null, "replies": [] }, { "id": 322548, "author_name": "asalzburger", "author_url": "", "post_date": "05/03/2018 07:04:59", "content": "

Indeed - below 3 hits there's simply not enough degrees of freedom to constrain the trajectory.

", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "322453": "The awards tab mentions: “...with respect to the traditional particle physics combinatorial approach...”. As this appears to be crucial to final ranking, can you point to a (hopefully brief) turtorial or description of the combinatorial approach?", "322460": "Hey pete, check out [page 20 of this talk][1]. I found it via [this Kaggle discussion][2], posted by David Rousseau, one of the challenge creators.\n\n\n [1]: https://drive.google.com/file/d/1tW0tSJAgPJlQOwoSQPG7mtzDd-08F4z1/view\n [2]: https://www.kaggle.com/c/trackml-particle-identification/discussion/55726#321441", "322477": "I read that the organisers will release basline code for hough transform and dbscan later. Let me know if i am wrong.", "322533": "The applied method in the two main experiment at the LHC is conceptional very simple:\n- build track seeds from points of three\n- run a (combinatorial) Kalman Filter to complete your trajectories\n\nQuite lengthy, but here is some background information:\n\n Cornelissen, T. et al., *Concepts, Design and Implementation of the ATLAS New Tracking (NEWT)*, ATL-SOFT-PUB-2007-007, https://cds.cern.ch/record/1020106\n\nAdam, W. et al., *Track reconstruction in the CMS tracker*, CERN-CMS-NOTE-2006-041, CMS-NOTE-2006-041", "322545": "Thanks - very helpful.", "322546": "OK that makes sense. I guess you take every “sensible” three point combinations (that could possible be in a track) and run the Kalman filter on it.", "322548": "Indeed - below 3 hits there's simply not enough degrees of freedom to constrain the trajectory." }, "source": "meta" }