MLS-C01 · Question #262
A credit card company wants to identify fraudulent transactions in real time. A data scientist builds a machine learning model for this purpose. The transactional data is captured and stored in…
The correct answer is A. Apply the Synthetic Minority Oversampling Technique (SMOTE) on the minority class in the. Given a fraud detection model with a high False Negative Rate, indicating a failure to detect many positive (fraud) cases, applying SMOTE on the minority class will best improve the model's performance by addressing class imbalance.
Question
A credit card company wants to identify fraudulent transactions in real time. A data scientist builds a machine learning model for this purpose. The transactional data is captured and stored in Amazon S3. The historic data is already labeled with two classes: fraud (positive) and fair transactions (negative). The data scientist removes all the missing data and builds a classifier by using the XGBoost algorithm in Amazon SageMaker. The model produces the following results:
- True positive rate (TPR): 0.700
- False negative rate (FNR): 0.300
- True negative rate (TNR): 0.977
- False positive rate (FPR): 0.023
- Overall accuracy: 0.949
Which solution should the data scientist use to improve the performance of the model?
Options
- AApply the Synthetic Minority Oversampling Technique (SMOTE) on the minority class in the
- BApply the Synthetic Minority Oversampling Technique (SMOTE) on the majority class in the
- CUndersample the minority class.
- DOversample the majority class.
How the community answered
(38 responses)- A74% (28)
- B13% (5)
- C3% (1)
- D11% (4)
Why each option
Given a fraud detection model with a high False Negative Rate, indicating a failure to detect many positive (fraud) cases, applying SMOTE on the minority class will best improve the model's performance by addressing class imbalance.
The provided metrics show a True Negative Rate (TNR) of 0.977 (good at identifying fair transactions) but a False Negative Rate (FNR) of 0.300 (poor at identifying fraud), which is characteristic of an imbalanced dataset where the fraud (positive) class is the minority. Applying the Synthetic Minority Oversampling Technique (SMOTE) on the minority class generates synthetic examples, helping the model learn better representations of the fraud class and improving its detection capability.
Applying SMOTE on the majority class would further increase the imbalance, making the model even more biased towards the majority class and worsening fraud detection.
Undersampling the minority class would reduce the already limited data for fraud cases, making it even harder for the model to learn and detect fraudulent transactions.
Oversampling the majority class would exacerbate the class imbalance, leading to a model that is more biased towards predicting non-fraudulent transactions, thus missing more actual fraud cases.
Concept tested: Class imbalance, SMOTE, fraud detection metrics
Source: https://imbalanced-learn.org/stable/references/generated/imblearn.over_sampling.SMOTE.html
Topics
Community Discussion
No community discussion yet for this question.