Prediction Scatter Plot

🔮 What is this?

This is the report card for PLS models. It compares the model's predicted values (what I guess) with the actual values (what it really is).

It allows you to see at a glance: How accurate is the model?

🧐 How to read?

• X-axis: Actual observed values.
• Y-axis: Model predicted values.
• Ideal line: There is a 45° diagonal line.
• Good model: All points neatly line up on this line.
• Bad model: Points are scattered like beans everywhere, completely unrelated.

🛠️ How to use?

Note: This chart has a lasso function for selecting abnormal parameters.

• Check distribution: If points are scattered, it indicates the model has large errors (high RMSE).
• Find deviations: If points are generally below the line, it indicates the model has systematic underestimation (Bias).