Which of the following is a simple statistic to monitor for categorical feature drift?
Which of the following is a simple statistic to monitor for categorical feature drift?
The most appropriate simple statistic to monitor for categorical feature drift is mode. Mode represents the most frequent category and can easily show any changes or shifts in the distribution of categorical data. While the other options provide useful information, they do not directly indicate categorical feature drift as effectively as mode.
would go with C
Among the options, the most appropriate simple statistic to monitor for categorical feature drift is mode. The others aren't right due to the below reasons: Number of unique values - while useful for understanding the diversity of categories, it does not directory indicate drive unless the number of categories changes significantly. Percentage of missing values - while important for data quality, it does not directly indicate drift in the distribution of categorical values.