Learning from Kaggling : Analyses of Grupo Bimbo inventory demand and Rossmann Store sales Competitions.
碩士 === 國立東華大學 === 應用數學系 === 105 === Kaggle is a well-known and very active machine learning analysis competition platform. In this study, we analyze the formerly featured competition, Grupo Bimbo inventory demand and Rossmann Store sales. Built upon XGBoost, our fitted model has achieved top 10% amo...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Published: |
2017
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/06394235884337178801 |
| Summary: | 碩士 === 國立東華大學 === 應用數學系 === 105 === Kaggle is a well-known and very active machine learning analysis competition platform. In this study, we analyze the formerly featured competition, Grupo Bimbo inventory demand and Rossmann Store sales.
Built upon XGBoost, our fitted model has achieved top 10% among all competing machines compared with the private leaderboard scores.
The preparation and data import, variable selection as well as feature manufacturing all have marked impact on the resultant performance of the fitted model. We will discuss these issues and share our experience in Kaggling.
More at https://github.com/f496328mm
|
|---|