Exam DP-100 topic 1 question 15 discussion

As there is a lack of information regarding the dataset in question, utilizing Custom Substitution would be advantageous. This allows for greater control in selecting appropriate substitutions that align with the data, resulting in a favourable outcome. Therefore, the answer is yes.

See latest reference document at here https://learn.microsoft.com/en-us/azure/machine-learning/component-reference/clean-missing-data?view=azureml-api-2#:~:text=Custom%20substitution%20value%3A%20Use%20this%20option%20to%20specify%20a%20placeholder%20value Quote 'For Cleaning Mode, select one of the following options for replacing or removing missing values: Custom substitution value: Use this option to specify a placeholder value (such as a 0 or NA) that applies to all missing values. The value that you specify as a replacement must be compatible with the data type of the column.' Yes. We can use 'Custom substitution value'.

https://learn.microsoft.com/en-us/previous-versions/azure/machine-learning/studio-module-reference/clean-missing-data#bkmk_ReplaceMissing - Replace using MICE (Multivariate Imputation using Chained Equations) - Custom substitution value (*) - Replace with mean - Replace with median: Calculates the column median value, and uses the median value as the replacement for any missing value in the column. - Replace with mode - Remove entire row - Replace using Probabilistic PCA 'rows with null and missing values', so no removing entire row. I choose YES (*)

If the requirement is to effectively handle missing values in a way that preserves the integrity of the dataset for machine learning purposes, you might need to consider other options as well, such as imputation methods (mean, median, mode, MICE) that can maintain statistical properties of the dataset. The best method depends on the nature of your data and the specific requirements of your machine learning model.

Answer B: All the replace missing value options apply to the column. The question talks about randomly missing values in the row. Hence removing the row will be ideal.

You dont know the dataset and therefore you cant assume A is correct. This means B is correct in this case. The only option that could be correct withoud seing the dataset is to remove the rows when missing values exist.

I think the tricky part is the null values. Clean missing data lets you input for that but doesn't address the null or NaN values.

Answer is 'A', here is the reference for those who have doubts: https://learn.microsoft.com/en-us/azure/machine-learning/component-reference/clean-missing-data

If the columns have different data types, we can not set one custom substitution for all the missing values. The only strategy that can apply to all data types is the mode.

yes the answer is A.

Custom substitution value: Use this option to specify a placeholder value (such as a 0 or NA) that applies to all missing values. The value that you specify as a replacement must be compatible with the data type of the column. https://learn.microsoft.com/en-us/azure/machine-learning/component-reference/clean-missing-data

Should be A

A is correct, Clean Missing Data module also provides "Custom substitution value" cleaning mode

Since it says custom, the answer would be Yes. A

answer should be YES, because we can use '0' for numeric and 'na' for text columns