Why should you care?

AIPredict media demo trains teams on explicit stars before implicit e-commerce clicks.

See it live — copy this example

Use a .NET console or Web API project with Microsoft.ML. Run dotnet run after pasting.

public class WatchRow { public uint UserId; public uint MovieId; public float Stars; }
var pipe = ml.Transforms.Conversion.MapValueToKey("U", nameof(WatchRow.UserId))
    .Append(ml.Transforms.Conversion.MapValueToKey("M", nameof(WatchRow.MovieId)))
    .Append(ml.Recommendation().Trainers.MatrixFactorization(new MatrixFactorizationTrainer.Options {
        MatrixColumnIndexColumnName = "U", MatrixRowIndexColumnName = "M", LabelColumnName = nameof(WatchRow.Stars)
    }));
var model = pipe.Fit(watches);

What happened?

• Explicit star labels fit MF cleanly.
• Evaluate with ranking metrics on held-out user–movie pairs.

Practice next

1. Load MovieLens-style CSV.
2. Train MF on Stars.
3. Predict top movies per user.
4. Switch to implicit play events.
5. Measure NDCG on holdout.

Remember

Same MF pattern. Explicit labels. Rank for display.

AIPredict media lab

Team practices recs on movie stars.

Outcome: Same pipeline ports to ShopNest clicks.