I have a requirement to recommend validation rules (from a set of pre-defined rules library) based on the column selected. I would like to understand the pattern of the data for the particular column and suggest rules accordingly.
Example: if the column entered has dates as values. My system should understand that it has to suggest all rules relevant for validating dates: Rules from rule lib like (dare format, date should lie between 1900-current, start date< end date etc..)
What kind if approach can I follow? -Recommendation engine
- topic similarity -anyother, combination of models?
Any ideas and suggestions would help