microsoft/SynapseML

[LightGBM] Regression: how to penalize negative predictions

Open

#918 建立於 2020年9月2日

在 GitHub 查看
 (18 留言) (0 反應) (2 負責人)Scala (5,228 star) (861 fork)batch import
area/lightgbmenhancementgood first issue

描述

I have a simple regression task (using a LightGBMRegressor) where I want to penalize negative predictions more than positive ones. Is there a way to achieve this with the default regression LightGBM objectives (see https://lightgbm.readthedocs.io/en/latest/Parameters.html)? If not, is it somehow possible to define (many example for default LightGBM model) and pass a custom regression objective?

Thx in advance!

貢獻者指南

技術棧
scala
領域
machine learning
議題類型
feature
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
half day
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
blocked
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
clear
前置要求
LightGBM basicsCustom objective functions
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
40
研究方向
Investigate LightGBM's documentation on custom objectives (https://lightgbm.readthedocs.io). Check the SynapseML source code for the LightGBM integration (likely in the 'lightgbm' folder). Look for any existing custom objective examples in the repository's issues or PRs. Engage with the assignees to understand if they have a proposed solution.
[LightGBM] Regression: how to penalize negative predictions · microsoft/SynapseML#918 | Good First Issue