ensemble

Video Presentation: Multi-variate (15,000) Time-Series Forecasting for Inventory Items

Recently, I was working on a project while using bagging tree-based algorithms for low-frequency time-series data. Originally, the data set contained only three variables for each series. However, given the nature of time-series data stored in the wide formar, I succeeded in engineeing 300 more features. The presentation was first introduced during an online session organized by the Aggregate Intelect Socratic Circles https://aisc.ai.science/