BP5 had the highest prediction error (MAE 0.080) — it was the hardest property for the model to learn, likely due to a more complex or non-linear relationship with the component inputs. BP8 was the second hardest (MAE 0.055). All other properties achieved MAE below 0.045.
BP1 · R²
0.998
Excellent
BP2 · R²
0.998
Excellent
BP3 · R²
0.998
Excellent
BP4 · R²
0.997
Excellent
BP5 · R²
0.946
Good
BP6 · R²
0.998
Excellent
BP7 · R²
0.997
Excellent
BP8 · R²
0.994
Excellent
BP9 · R²
0.997
Excellent
BP10 · R²
0.998
Excellent
How accurate is the model?
The scatter plot below compares what the model predicted vs what the value actually was for 500 fuel blend samples. Each dot is one sample. If the model were perfect, every dot would sit on the diagonal line. The tighter the dots cluster around that line, the better. Click any BP card above or use the buttons to switch between blend properties.
Predicted vs Actual · Scatter
Blend Property 1
FIT ACCURACY
R² — How well did the model do per property?
R² (R-squared) is a score from 0 to 1. A score of 1.0 means the model predicted every value perfectly. A score of 0.998 means it got 99.8% of the way there — near perfect. BP5 at 0.946 is the hardest property for the model to predict.
R² Score by Property
MODEL PERFORMANCE
MAE — How far off were the predictions on average?
MAE (Mean Absolute Error) is the average gap between predicted and actual values. Lower is better. Most properties have an MAE around 0.036–0.055, meaning predictions are typically off by less than 0.05 units. BP5 has the highest error at 0.08.
Mean Absolute Error by Property
MAE BREAKDOWN
Radar — Predicted vs Actual for one sample
This web chart shows all 10 blend properties at once for a single fuel blend sample. The orange shape is what the model predicted. The blue dashed shape is the actual measured value. The closer the two shapes overlap, the better the prediction.
Prediction vs Actual · All Properties · Sample
SAMPLE RADAR