An unpruned decision tree achieves 99% accuracy on training data but only 60% on test data. What is this phenomenon called?

Question

BlackTom AI · Accepted Answer

Overfitting occurs when a model fits the training data too closely, capturing noise or specific idiosyncrasies rather than the underlying pattern. That leads to excellent training performance but poor generalization to new (test) data.

1. Relevant concepts:
- Training accuracy: how well the model predicts examples it was trained on.
- Test accuracy: how well it predicts unseen data; proxy for generalization.
- Overfitting: high variance, low bias; model complexity too high relative to available signal.
- Underfitting: high bias, model too simple to capture patterns.
- Bias–variance tradeoff: balancing model complexity to minimize total error.
- Pruning (for decision trees): reduces complexity to improve generalization.

2. Step-by-step reasoning:
- Observed numbers: training accuracy = 99%, test accuracy = 60%.
- A gap of 39 percentage points is large; model performs almost perfectly on training but poorly on unseen data.
- A decision tree that is unpruned can grow very deep and create leaves that memorize training instances (noise).
- Memorization yields near-100% training accuracy but fails to generalize, which is the hallmark of overfitting.

3. Calculations / intermediate reasoning:
- Difference = 99% − 60% = 39% gap indicating substantial generalization error.
- Such a gap implies variance-dominated error (model complexity too high).

4. Why the correct answer is correct:
- Overfitting directly describes this situation: excessive model complexity (unpruned tree) causing near-perfect training results but poor test performance. Pruning or regularization are standard remedies.

5. Why other options are incorrect:
- Underfitting: would show low training and low test accuracy (model too simple), not high training accuracy.
- Bias–variance tradeoff: this is a broader concept describing the underlying forces, not the specific phenomenon observed. The observed case is the overfitting side of that tradeoff.
- Cross-validation failure: refers to problems in the evaluation procedure, not the model’s characteristic of fitting noise. The metrics given already show the model’s behavior, not a validation process error.

6. Conclusion:
The described phenomenon is overfitting, which matches the provided answer: Overfitting.

An unpruned decision tree achieves 99% accuracy on training data but only 60% on test data. What is this phenomenon called?单项选择题

类似问题

In the above image, the prediction made at any value of X is shown by the blue line. This predictive model is an overfit for the training data.

Which one of the following linear discriminants is most prone to overfitting a training data set?

Overfitting (1 mark)What is overfitting?[Fill in the blank]

Which of the following conditions can lead to logistic regression overfitting the training data?

Question at position 6 True/False Question: There is no direct relationship between model complexity and overfitting. TrueFalse

更多留学生实用工具

智能学习助手

风格化写作助手

论文查重助手

文献引用助手

课堂转译助手

课堂笔记助手

Quiz搜索助手

学校历年真题

智能刷题助手

智能匹配练习

希望你的学习变得更简单