Thank you to everyone for participating in our MLB Player Digital Engagement Forecasting Kaggle Competition. We at the MLB ML team were thrilled to collaborate with this community and Kaggle to support machine learning and exploring more ways to better connect fans with the great game of baseball.
\nWe wanted to include an Explainability Prize in addition to the Leaderboard Prize so that we could hear more directly from the creators of these submissions. Machine Learning development is rarely linear, so getting some broader input on how these were created we thought would be helpful to us as well as the competitors.
\nPlease join us in congratulating yingzhou0510 (1st Place), ryosukesaito (2nd Place) and hijest (3rd Place) for their top submissions.
\nWell done! This entry clearly describes the problem space and explains the methodology for employing a random forest to arrive at a solution. This type of random forest strategy is one that a data scientist could use in the real world.
\nThis entry’s code quality is top-notch. The step-by-step explanation is by far the best in the competition. The way this entry explains how they utilize lag features within time series data and MSE for dealing with outliers is very thorough.
\nThe way the data is explored, profiled and visualized in this entry is precisely how a data scientist should approach an unknown dataset. The iterative nature with which this entry explains feature selection and feature engineering is outstanding.
", "messages": [ { "id": 1535401, "postDate": "2021-10-05T19:07:30.913Z", "content": "Thank you to everyone for participating in our MLB Player Digital Engagement Forecasting Kaggle Competition. We at the MLB ML team were thrilled to collaborate with this community and Kaggle to support machine learning and exploring more ways to better connect fans with the great game of baseball.
\nWe wanted to include an Explainability Prize in addition to the Leaderboard Prize so that we could hear more directly from the creators of these submissions. Machine Learning development is rarely linear, so getting some broader input on how these were created we thought would be helpful to us as well as the competitors.
\nPlease join us in congratulating yingzhou0510 (1st Place), ryosukesaito (2nd Place) and hijest (3rd Place) for their top submissions.
\nWell done! This entry clearly describes the problem space and explains the methodology for employing a random forest to arrive at a solution. This type of random forest strategy is one that a data scientist could use in the real world.
\nThis entry’s code quality is top-notch. The step-by-step explanation is by far the best in the competition. The way this entry explains how they utilize lag features within time series data and MSE for dealing with outliers is very thorough.
\nThe way the data is explored, profiled and visualized in this entry is precisely how a data scientist should approach an unknown dataset. The iterative nature with which this entry explains feature selection and feature engineering is outstanding.
", "rawMarkdown": "####***From Major League Baseball***####\n\nThank you to everyone for participating in our [MLB Player Digital Engagement Forecasting Kaggle Competition](https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting). We at the MLB ML team were thrilled to collaborate with this community and Kaggle to support machine learning and exploring more ways to better connect fans with the great game of baseball.\n\nWe wanted to include an Explainability Prize in addition to the Leaderboard Prize so that we could hear more directly from the creators of these submissions. Machine Learning development is rarely linear, so getting some broader input on how these were created we thought would be helpful to us as well as the competitors.\n\nPlease join us in congratulating [**yingzhou0510**](https://www.kaggle.com/yingzhou0510) (1st Place), [**ryosukesaito**](https://www.kaggle.com/ryosukesaito) (2nd Place) and [**hijest**](https://www.kaggle.com/hijest) (3rd Place) for their top submissions.\n\n\n## First Place: [Explanatory Analysis Using Random Forest](https://www.kaggle.com/yingzhou0510/explanatory-analysis-using-random-forest) ##\n\nWell done! This entry clearly describes the problem space and explains the methodology for employing a random forest to arrive at a solution. This type of random forest strategy is one that a data scientist could use in the real world.\n\n\n## Second Place: [Insights of MLB engagement (lag / MSE)](https://www.kaggle.com/ryosukesaito/insights-of-mlb-engagement-lag-mse) ##\n\nThis entry’s code quality is top-notch. The step-by-step explanation is by far the best in the competition. The way this entry explains how they utilize lag features within time series data and MSE for dealing with outliers is very thorough.\n\n\n## Third Place: [Predict Player Engagement large EDA](https://www.kaggle.com/hijest/predict-player-engagement-large-eda) ##\n\nThe way the data is explored, profiled and visualized in this entry is precisely how a data scientist should approach an unknown dataset. The iterative nature with which this entry explains feature selection and feature engineering is outstanding.", "votes": 12 }, { "id": 1537654, "postDate": "2021-10-07T16:30:02.570Z", "rawMarkdown": "", "votes": 1, "isDeleted": true } ], "comments": [ { "id": 1537654, "author_name": "", "author_url": "", "post_date": "2021-10-07T16:30:02.570000", "content": "", "votes": 1, "replies": [] } ], "raw_markdown_by_id": { "1535401": "####***From Major League Baseball***####\n\nThank you to everyone for participating in our [MLB Player Digital Engagement Forecasting Kaggle Competition](https://www.kaggle.com/c/mlb-player-digital-engagement-forecasting). We at the MLB ML team were thrilled to collaborate with this community and Kaggle to support machine learning and exploring more ways to better connect fans with the great game of baseball.\n\nWe wanted to include an Explainability Prize in addition to the Leaderboard Prize so that we could hear more directly from the creators of these submissions. Machine Learning development is rarely linear, so getting some broader input on how these were created we thought would be helpful to us as well as the competitors.\n\nPlease join us in congratulating [**yingzhou0510**](https://www.kaggle.com/yingzhou0510) (1st Place), [**ryosukesaito**](https://www.kaggle.com/ryosukesaito) (2nd Place) and [**hijest**](https://www.kaggle.com/hijest) (3rd Place) for their top submissions.\n\n\n## First Place: [Explanatory Analysis Using Random Forest](https://www.kaggle.com/yingzhou0510/explanatory-analysis-using-random-forest) ##\n\nWell done! This entry clearly describes the problem space and explains the methodology for employing a random forest to arrive at a solution. This type of random forest strategy is one that a data scientist could use in the real world.\n\n\n## Second Place: [Insights of MLB engagement (lag / MSE)](https://www.kaggle.com/ryosukesaito/insights-of-mlb-engagement-lag-mse) ##\n\nThis entry’s code quality is top-notch. The step-by-step explanation is by far the best in the competition. The way this entry explains how they utilize lag features within time series data and MSE for dealing with outliers is very thorough.\n\n\n## Third Place: [Predict Player Engagement large EDA](https://www.kaggle.com/hijest/predict-player-engagement-large-eda) ##\n\nThe way the data is explored, profiled and visualized in this entry is precisely how a data scientist should approach an unknown dataset. The iterative nature with which this entry explains feature selection and feature engineering is outstanding.", "1537654": "" } }