{ "id": 97664, "title": "How is the origin chosen for Cartesian coordinates?", "url": "/competitions/champs-scalar-coupling/discussion/97664", "author_name": "", "post_date": "2019-06-28T08:54:22.251185900Z", "votes": 3, "comment_count": 3, "views": 0, "content": "

Given the Cartesian coordinates of each atom in a molecule, does the origin (0, 0, 0) in the system have a meaning? What is the convention/ reference point used for assigning the x y z values to each atom in a given molecule? Can someone clarify this point?

", "messages": [ { "id": "563418", "postDate": "06/28/2019 08:54:22", "content": "

Given the Cartesian coordinates of each atom in a molecule, does the origin (0, 0, 0) in the system have a meaning? What is the convention/ reference point used for assigning the x y z values to each atom in a given molecule? Can someone clarify this point?

", "rawMarkdown": "Given the Cartesian coordinates of each atom in a molecule, does the origin (0, 0, 0) in the system have a meaning? What is the convention/ reference point used for assigning the x y z values to each atom in a given molecule? Can someone clarify this point?", "votes": null }, { "id": "563823", "postDate": "06/28/2019 18:07:25", "content": "

It's arbitrary. You should not use absolute coordinates (X, Y, Z). You should only use features based on how one atom is positioned relative to another (distance, angle, etc).

", "rawMarkdown": "It's arbitrary. You should not use absolute coordinates (X, Y, Z). You should only use features based on how one atom is positioned relative to another (distance, angle, etc).", "votes": null }, { "id": "564212", "postDate": "06/29/2019 06:18:03", "content": "

For some examples of relative features see these kernels:\nhttps://www.kaggle.com/artgor/brute-force-feature-engineering\nhttps://www.kaggle.com/robertburbidge/distance-features\nhttps://www.kaggle.com/adrianoavelar/bond-calculaltion-lb-0-82

", "rawMarkdown": "For some examples of relative features see these kernels:\n[https://www.kaggle.com/artgor/brute-force-feature-engineering](https://www.kaggle.com/artgor/brute-force-feature-engineering)\n[https://www.kaggle.com/robertburbidge/distance-features](https://www.kaggle.com/robertburbidge/distance-features)\n[https://www.kaggle.com/adrianoavelar/bond-calculaltion-lb-0-82](https://www.kaggle.com/adrianoavelar/bond-calculaltion-lb-0-82)", "votes": null }, { "id": "564213", "postDate": "06/29/2019 06:20:36", "content": "

On a related point, the magnetic shielding tensors are given in the same frame of reference as the atom coordinates, but again, you shouldn't use the absolute values (which is a mistake I made). See the comments in this kernel for a discussion:\nhttps://www.kaggle.com/robertburbidge/evaluating-magnetic-shield-tensors

", "rawMarkdown": "On a related point, the magnetic shielding tensors are given in the same frame of reference as the atom coordinates, but again, you shouldn't use the absolute values (which is a mistake I made). See the comments in this kernel for a discussion:\n[https://www.kaggle.com/robertburbidge/evaluating-magnetic-shield-tensors](https://www.kaggle.com/robertburbidge/evaluating-magnetic-shield-tensors)", "votes": null } ], "comments": [ { "id": 563823, "author_name": "petersk20", "author_url": "", "post_date": "06/28/2019 18:07:25", "content": "

It's arbitrary. You should not use absolute coordinates (X, Y, Z). You should only use features based on how one atom is positioned relative to another (distance, angle, etc).

", "votes": null, "replies": [] }, { "id": 564212, "author_name": "robertburbidge", "author_url": "", "post_date": "06/29/2019 06:18:03", "content": "

For some examples of relative features see these kernels:\nhttps://www.kaggle.com/artgor/brute-force-feature-engineering\nhttps://www.kaggle.com/robertburbidge/distance-features\nhttps://www.kaggle.com/adrianoavelar/bond-calculaltion-lb-0-82

", "votes": null, "replies": [] }, { "id": 564213, "author_name": "robertburbidge", "author_url": "", "post_date": "06/29/2019 06:20:36", "content": "

On a related point, the magnetic shielding tensors are given in the same frame of reference as the atom coordinates, but again, you shouldn't use the absolute values (which is a mistake I made). See the comments in this kernel for a discussion:\nhttps://www.kaggle.com/robertburbidge/evaluating-magnetic-shield-tensors

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "563418": "Given the Cartesian coordinates of each atom in a molecule, does the origin (0, 0, 0) in the system have a meaning? What is the convention/ reference point used for assigning the x y z values to each atom in a given molecule? Can someone clarify this point?", "563823": "It's arbitrary. You should not use absolute coordinates (X, Y, Z). You should only use features based on how one atom is positioned relative to another (distance, angle, etc).", "564212": "For some examples of relative features see these kernels:\n[https://www.kaggle.com/artgor/brute-force-feature-engineering](https://www.kaggle.com/artgor/brute-force-feature-engineering)\n[https://www.kaggle.com/robertburbidge/distance-features](https://www.kaggle.com/robertburbidge/distance-features)\n[https://www.kaggle.com/adrianoavelar/bond-calculaltion-lb-0-82](https://www.kaggle.com/adrianoavelar/bond-calculaltion-lb-0-82)", "564213": "On a related point, the magnetic shielding tensors are given in the same frame of reference as the atom coordinates, but again, you shouldn't use the absolute values (which is a mistake I made). See the comments in this kernel for a discussion:\n[https://www.kaggle.com/robertburbidge/evaluating-magnetic-shield-tensors](https://www.kaggle.com/robertburbidge/evaluating-magnetic-shield-tensors)" }, "source": "meta" }