In many of the notebooks, people are using a Pinball loss function for training the model. Can somebody explain the use of this loss fuction along with the evaluation metrics commonly used for this type of problems?
", "messages": [ { "id": "984020", "postDate": "08/24/2020 19:17:05", "content": "In many of the notebooks, people are using a Pinball loss function for training the model. Can somebody explain the use of this loss fuction along with the evaluation metrics commonly used for this type of problems?
", "rawMarkdown": "In many of the notebooks, people are using a Pinball loss function for training the model. Can somebody explain the use of this loss fuction along with the evaluation metrics commonly used for this type of problems?", "votes": null }, { "id": "984027", "postDate": "08/24/2020 19:20:14", "content": "Yeah! and also do tell me about how to calculate the confidence. Thanks for your input
", "rawMarkdown": "Yeah! and also do tell me about how to calculate the confidence. Thanks for your input", "votes": null }, { "id": "984297", "postDate": "08/25/2020 02:57:31", "content": "https://www.lokad.com/pinball-loss-function-definition
\nI think people are using Quantile Regression to calculate the confidence in the FVC (using the 0.2nd and 0.8th quantiles for example) and the pin ball loss function is basically looks like a loss measure for this kind of regression.
Thank you @jonykarki.
", "rawMarkdown": "Thank you @jonykarki.", "votes": null }, { "id": "1003771", "postDate": "09/09/2020 08:29:41", "content": "The pinball loss function is a metric used to assess the accuracy of a quantile forecast.
\nEvaluating the accuracy of a quantile forecast is a subtle problem. Indeed, contrary to the classic forecasts where the goal is to have the forecast as close as possible from the observed values, the situation is biased (on purpose) when it comes to quantile forecasts. Hence the naive comparison observed vs forecasts is not satisfying. The pinball loss function returns a value that can be interpreted as the accuracy of a quantile forecasting model.
", "rawMarkdown": "The pinball loss function is a metric used to assess the accuracy of a quantile forecast.\n\nEvaluating the accuracy of a quantile forecast is a subtle problem. Indeed, contrary to the classic forecasts where the goal is to have the forecast as close as possible from the observed values, the situation is biased (on purpose) when it comes to quantile forecasts. Hence the naive comparison observed vs forecasts is not satisfying. The pinball loss function returns a value that can be interpreted as the accuracy of a quantile forecasting model.", "votes": null } ], "comments": [ { "id": 984027, "author_name": "itsjackthereaper", "author_url": "", "post_date": "08/24/2020 19:20:14", "content": "Yeah! and also do tell me about how to calculate the confidence. Thanks for your input
", "votes": null, "replies": [] }, { "id": 984297, "author_name": "jonykarki", "author_url": "", "post_date": "08/25/2020 02:57:31", "content": "https://www.lokad.com/pinball-loss-function-definition
\nI think people are using Quantile Regression to calculate the confidence in the FVC (using the 0.2nd and 0.8th quantiles for example) and the pin ball loss function is basically looks like a loss measure for this kind of regression.
Thank you @jonykarki.
", "votes": null, "replies": [] } ] }, { "id": 1003771, "author_name": "avikdas29", "author_url": "", "post_date": "09/09/2020 08:29:41", "content": "The pinball loss function is a metric used to assess the accuracy of a quantile forecast.
\nEvaluating the accuracy of a quantile forecast is a subtle problem. Indeed, contrary to the classic forecasts where the goal is to have the forecast as close as possible from the observed values, the situation is biased (on purpose) when it comes to quantile forecasts. Hence the naive comparison observed vs forecasts is not satisfying. The pinball loss function returns a value that can be interpreted as the accuracy of a quantile forecasting model.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "984020": "In many of the notebooks, people are using a Pinball loss function for training the model. Can somebody explain the use of this loss fuction along with the evaluation metrics commonly used for this type of problems?", "984027": "Yeah! and also do tell me about how to calculate the confidence. Thanks for your input", "984297": "https://www.lokad.com/pinball-loss-function-definition\nI think people are using Quantile Regression to calculate the confidence in the FVC (using the 0.2nd and 0.8th quantiles for example) and the pin ball loss function is basically looks like a loss measure for this kind of regression.", "991761": "Thank you @jonykarki.", "1003771": "The pinball loss function is a metric used to assess the accuracy of a quantile forecast.\n\nEvaluating the accuracy of a quantile forecast is a subtle problem. Indeed, contrary to the classic forecasts where the goal is to have the forecast as close as possible from the observed values, the situation is biased (on purpose) when it comes to quantile forecasts. Hence the naive comparison observed vs forecasts is not satisfying. The pinball loss function returns a value that can be interpreted as the accuracy of a quantile forecasting model." }, "source": "meta" }