The correct answer is A. low variance.
Leave-one-out cross-validation (LOOCV) is a resampling method for estimating the error rate of a model on unseen data. It is a special case of k-fold cross-validation where k=n, the number of data points. In LOOCV, the model is trained on all data points except one, and then the model is evaluated on the left-out data point. This process is repeated for each data point, and the average error rate is reported.
LOOCV is often considered to be the most accurate method for estimating the error rate of a model. However, it is also the most computationally expensive method. This is because the model must be trained and evaluated n times, once for each data point.
The variance of a model is a measure of how much the model’s predictions change when the training data is changed. A model with low variance is less sensitive to changes in the training data. This is important because we want our model to generalize well to unseen data.
LOOCV can lead to a model with low variance. This is because the model is trained on all data points except one, and then the model is evaluated on the left-out data point. This means that the model is only evaluated on a single data point at a time. This can lead to the model overfitting the training data, and therefore having low variance.
In conclusion, the major issue in Leave-One-Out-Cross-Validation(LOOCV) is low variance. This is because LOOCV can lead to the model overfitting the training data.
Here is a brief explanation of each option:
- A. low variance. LOOCV can lead to a model with low variance. This is because the model is trained on all data points except one, and then the model is evaluated on the left-out data point. This means that the model is only evaluated on a single data point at a time. This can lead to the model overfitting the training data, and therefore having low variance.
- B. high variance. LOOCV does not necessarily lead to a model with high variance. In fact, LOOCV can lead to a model with low variance, as discussed above.
- C. faster runtime compared to k-fold cross validation. LOOCV is the most computationally expensive method for estimating the error rate of a model. This is because the model must be trained and evaluated n times, once for each data point.
- D. slower runtime compared to normal validation. LOOCV is not necessarily slower than normal validation. In fact, LOOCV can be faster than normal validation if the model is trained using a technique such as stochastic gradient descent.