From the course: Strategic Data Collection, Modeling, and Quality Management for AI Systems
Join today to access over 24,900 courses taught by industry experts.
Improving outcomes with data
From the course: Strategic Data Collection, Modeling, and Quality Management for AI Systems
Improving outcomes with data
- [Narrator] Let's talk about how to transform good data into great results. You might have heard the phrase GIGO or garbage in, garbage out but the reverse is also true. High quality data properly prepared and maintain can dramatically improve your model's performance. Consider a retail forecasting system that was able to improve its accuracy from 65% to 89% through three key data optimization strategies. These optimization strategies are feature engineering, dataset balancing, and continuous updates. Let's examine each one starting with feature engineering. First, feature engineering the art of creating better input variables. Raw data rarely tells the whole story. For example, instead of just using raw sales data, you might create features like rolling averages over different time periods, seasonality indicators, promotion, impact scores, and stock out risk indicators. Each engineered feature gives your model new ways of understanding patterns in the data. But here's the key. These…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.