Question 31

Question

Question 31

BlackTom AI · Accepted Answer

Accuracy measures the proportion of correct predictions: (TP + TN) / (TP + TN + FP + FN). Precision and recall are class-specific: precision = TP / (TP + FP) and recall = TP / (TP + FN). These definitions are the key knowledge points.

Step-by-step reasoning with a concrete example to show why accuracy can be misleading under class imbalance:
- Suppose a binary dataset has 100 examples: 95 negatives and 5 positives.
- A trivial classifier that predicts “negative” for every example yields TP = 0, TN = 95, FP = 0, FN = 5.
- Compute accuracy: (TP + TN) / total = (0 + 95) / 100 = 0.95 = 95%.
- Compute recall for the positive class: TP / (TP + FN) = 0 / (0 + 5) = 0%.
- Compute precision for the positive class: TP / (TP + FP) = 0 / (0 + 0) which is undefined or treated as 0 in practice.
This shows a model with 95% accuracy yet completely fails to detect the positive class (recall = 0). Thus accuracy alone hides poor performance on minority classes.

Why the correct answer is correct:
- Option d states accuracy alone can be misleading, especially when classes are imbalanced. The example above directly demonstrates this: high accuracy but useless behavior for the important minority class.

Why the other options are incorrect:
- Option a: Precision and recall are not always larger than accuracy — they are different metrics and can be higher, lower, or undefined depending on TP, FP, FN.
- Option b: Accuracy, precision and recall are evaluation metrics applicable to any dataset (training or test); they are not restricted to training vs test sets.
- Option c: Precision and recall do not remove outliers or extreme points; they measure correct/incorrect predictions for a given class.

Conclusion: Therefore the best choice is d. Accuracy alone can be misleading, especially when the classes are imbalanced.

Question 31单项选择题

类似问题

以下哪些是评价推荐系统准确率的方法？（多选题） Which of the following are ways to evaluate the accuracy of a recommender system? Select all that apply

医疗人工智能模型用于检测500名患者的疾病：实际上有100名患者患有这种疾病模型正确识别了其中70种它错误地将50名健康患者诊断为患病问题：计算模型的精度。计算型号的召回率。

Which metric answers the question: “Of the items the model said were positive, how many actually were positive?”

Sometimes we have to tune a model to bias more towards a certain type of error or performance metric (e.g., false positives, false negatives, precision, recall) because of the nature of a specific use case (e.g., cancer detection/prediction vs spam detection).

What was the world's first widely adopted biodiversity policy?

The aim of the Paris Agreement is to ensure that the earth’s preindustrial temperature is not exceeded by more than º Celsius.

The IPCC's mission is to Blank ______.

Why have most international agreements for environmental protection been based on the honor system?

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