Forecasting using machine learning methods
summary
Covers forecasting sales and machinery failures using practical ML workflows: start from a data sample, begin with simple models and iterate, plan for explainability and retraining after market shocks. Prototypes are built with Databricks on Azure (Jupyter notebooks, Spark) and benchmarking uses Azure Machine Learning AutoML; models tried include Prophet, linear regression and ensemble methods. Presents an NCBR research project producing an industrial early‑failure detector for an assembly screwdriver, and notes that simple regression gave measurable ROI in pilots.
AI-written summary · about
chapters (5)
- 0:00Wprowadzenie i firma
- 3:16Wyzwania prognozowania biznesowego
- 17:44Metody i technologie uczenia maszynowego
- 29:24Demo i warsztat inżyniera
- 36:50Podsumowanie i pytania
related talks
48mMachine Learning in Azure Databricks
39mForecasting With Classical and Machine Learning Methods
59mA developers guide to Machine Learning
26mPredicting the future (of equipment) using ML.NET
16mTransforming Inventory Forecasting with Production ML
53mMachine Learning in E‑commerce
21mHow The Application Of Machine Learning Is Changing The Way Companies Use Data
41mLeveraging Automated Machine Learning to enable anyone to develop Machine Learning solutions