VN1 Forecasting - Accuracy Challenge Phase 1
Share:
Finished
competition-bg

VN1 Forecasting - Accuracy Challenge Phase 1

Flieber, Syrup Tech, and SupChains Launch an AI-Driven Supply Chain Forecasting Competition

VN1 Forecasting - Accuracy Challenge
Machine Learning/AI
Enterprise
E-commerce/Retail
Total Prize 20,000
Scroll To Top

Olivier Sprangers · 18 September 2024

1052
Copied!
15

MLForecast starter

Uses LightGBM to generate predictions

Description

In this notebook, you'll find an application of LightGBM - a gradient-boosting machine model that has proven itself as one of the best forecasting methods (a.o. in the M5 forecasting competition). 

 

We go through the following steps:

  1. Importing the required packages
  2. Importing the data
  3. Feature engineering
  4. Model training
  5. Submission

There's tons of ways to improve this notebook - better feature engineering, better cross-validation, different loss functions, tune the hyperparameters, etc, but wanted to share for those who'd like to experiment with MLForecast methods.

 

Disclaimer: I work for Nixtla, authors of MLForecast.

      
      
    

Comments

avidar alkurdi

Posted 2 months ago
Great Notebook, I use the same library, just confused on this comment "There is implied target leakage in the Price feature - where price is unavailable, there is zero sales, and vice versa. So, we first lag the Price feature by one." This might be domain knowledge you might now, but how is that a data leak, are you trying to say price comes only comes after a sale? and why fill the price with the mean of the price and not the zero values or the mean of the uniquid specifically ?
Join our private community in Discord

Keep up to date by participating in our global community of data scientists and AI enthusiasts. We discuss the latest developments in data science competitions, new techniques for solving complex challenges, AI and machine learning models, and much more!