I had a question about the guidelines for the explainability prize and how it relates to what we might be exploring in our submission. The prompt asks us to study features which \"most contribute to digital engagement targets\", so this is certainly a wide range of things to study.
\nIn the evaluation criteria, there is an evaluation on whether the \"analyses can be applied to growing engagement measures\". Assuming this notebook is catered towards the MLB, it seems like the goal here is to, through some analysis, come up with ways MLB can improve engagement.
\nCertainly, there are many things that impact engagement completely out of the MLB's control (what players do off the field, what they post on social media, etc.). Should these be excluded from the study? Or should we simply analyze the features with the \"most contribution\", regardless of whether or not the MLB can act upon these features?
", "messages": [ { "id": "1364404", "postDate": "06/25/2021 00:21:59", "content": "I had a question about the guidelines for the explainability prize and how it relates to what we might be exploring in our submission. The prompt asks us to study features which \"most contribute to digital engagement targets\", so this is certainly a wide range of things to study.
\nIn the evaluation criteria, there is an evaluation on whether the \"analyses can be applied to growing engagement measures\". Assuming this notebook is catered towards the MLB, it seems like the goal here is to, through some analysis, come up with ways MLB can improve engagement.
\nCertainly, there are many things that impact engagement completely out of the MLB's control (what players do off the field, what they post on social media, etc.). Should these be excluded from the study? Or should we simply analyze the features with the \"most contribution\", regardless of whether or not the MLB can act upon these features?
", "rawMarkdown": "I had a question about the guidelines for the explainability prize and how it relates to what we might be exploring in our submission. The prompt asks us to study features which \"most contribute to digital engagement targets\", so this is certainly a wide range of things to study.\n\nIn the evaluation criteria, there is an evaluation on whether the \"analyses can be applied to growing engagement measures\". Assuming this notebook is catered towards the MLB, it seems like the goal here is to, through some analysis, come up with ways MLB can improve engagement. \n\nCertainly, there are many things that impact engagement completely out of the MLB's control (what players do off the field, what they post on social media, etc.). Should these be excluded from the study? Or should we simply analyze the features with the \"most contribution\", regardless of whether or not the MLB can act upon these features?", "votes": null }, { "id": "1364414", "postDate": "06/25/2021 00:42:29", "content": "Good questions! There are really no constraints here. Unlike in the code competition, where you are being measured against a pre-defined evaluation metric for your predictions' accuracy, the explainability prize was created to be a far more open-ended analysis. The key will be your ability to defend the measures you discover, backed by data. There is, however, a component of \"Application\" as you observed in the evaluation rubric which MLB will be considering in their selection for the prize. Keep in mind, however, that while there are behavioral attributes which MLB may not be able to control directly, it may be possible to influence to what extent those attributes are amplified or marketed, if they had the knowledge that those factors are, in fact, influential in engagement.
", "rawMarkdown": "Good questions! There are really no constraints here. Unlike in the code competition, where you are being measured against a pre-defined evaluation metric for your predictions' accuracy, the explainability prize was created to be a far more open-ended analysis. The key will be your ability to defend the measures you discover, backed by data. There is, however, a component of \"Application\" as you observed in the evaluation rubric which MLB will be considering in their selection for the prize. Keep in mind, however, that while there are behavioral attributes which MLB may not be able to control directly, it may be possible to influence to what extent those attributes are amplified or marketed, if they had the knowledge that those factors are, in fact, influential in engagement.", "votes": null } ], "comments": [ { "id": 1364414, "author_name": "juliaelliott", "author_url": "", "post_date": "06/25/2021 00:42:29", "content": "Good questions! There are really no constraints here. Unlike in the code competition, where you are being measured against a pre-defined evaluation metric for your predictions' accuracy, the explainability prize was created to be a far more open-ended analysis. The key will be your ability to defend the measures you discover, backed by data. There is, however, a component of \"Application\" as you observed in the evaluation rubric which MLB will be considering in their selection for the prize. Keep in mind, however, that while there are behavioral attributes which MLB may not be able to control directly, it may be possible to influence to what extent those attributes are amplified or marketed, if they had the knowledge that those factors are, in fact, influential in engagement.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1364404": "I had a question about the guidelines for the explainability prize and how it relates to what we might be exploring in our submission. The prompt asks us to study features which \"most contribute to digital engagement targets\", so this is certainly a wide range of things to study.\n\nIn the evaluation criteria, there is an evaluation on whether the \"analyses can be applied to growing engagement measures\". Assuming this notebook is catered towards the MLB, it seems like the goal here is to, through some analysis, come up with ways MLB can improve engagement. \n\nCertainly, there are many things that impact engagement completely out of the MLB's control (what players do off the field, what they post on social media, etc.). Should these be excluded from the study? Or should we simply analyze the features with the \"most contribution\", regardless of whether or not the MLB can act upon these features?", "1364414": "Good questions! There are really no constraints here. Unlike in the code competition, where you are being measured against a pre-defined evaluation metric for your predictions' accuracy, the explainability prize was created to be a far more open-ended analysis. The key will be your ability to defend the measures you discover, backed by data. There is, however, a component of \"Application\" as you observed in the evaluation rubric which MLB will be considering in their selection for the prize. Keep in mind, however, that while there are behavioral attributes which MLB may not be able to control directly, it may be possible to influence to what extent those attributes are amplified or marketed, if they had the knowledge that those factors are, in fact, influential in engagement." }, "source": "meta" }