Figure 9
The figure shows three line charts placed horizontally. The first chart is titled “Learning Curve: X G Boost”. The horizontal axis is labeled “Training examples” and ranges from 200 to 1000 in increments of 100 units. The vertical axis is labeled “Accuracy” and ranges from 0.90 to 1.00 in increments of 0.02 units. Two lines are shown, identified in the legend as “Train” and “Validation”. The training accuracy line stays constant at 1.00 across all training example values. The validation accuracy line starts at 0.89 at 100 training examples, increases slightly to 0.90 by 300 training examples, and then remains nearly flat through 1000 training examples. The second chart is titled “Learning Curve: CatBoost”. The horizontal axis is labeled “Training examples” and ranges from 200 to 1000 in increments of 100 units. The vertical axis is labeled “Accuracy” and ranges from 0.90 to 1.00 in increments of 0.02 units. Two lines are shown and identified in the legend as “Train” and “Validation”. The training accuracy line stays constant at 1.00 across all training example values. The validation accuracy line remains flat at 0.90 from 100 to 1000 training examples. The third chart is titled “Learning Curve: Light G B M”. The horizontal axis is labeled “Training examples” and ranges from 200 to 1000 in increments of 100 units. The vertical axis is labeled “Accuracy” and ranges from 0.90 to 1.00 in increments of 0.02 units. Two lines are displayed and identified in the legend as “Train” and “Validation”. The training accuracy line remains constant at 1.00 across all training example values. The validation accuracy line begins at approximately 0.89 at 100 training examples, increases slightly to around 0.90 by 300 training examples, and then stays nearly flat through 1000 training examples. Note: All numerical values are approximated.The comparison of learning-curve analysis between XGBoost, CatBoost, and LightGBM