What is the best method for estimating missing data?

Hey, everyone, I'm totally new in this issue in Matlab. I have 30 years of monthly precipitation data for a specific station, it has some missing data for some months. now I want to estimating this data. I search and found some function:

fillmissing and fillgaps. I don't know which is better in this case.

so please tell me what of them can be used in this case.

do you think any alternative way exists by the way? like ANN?

Thank you all

[Adam Danz]

There really isn't a one-size-fits-all best approach to filling in missing data and this problem really isn't related to Matlab (or python, excel, R, SPSS, you-name-it). Before digging into functions, you should think about it from a high perspective. What should that missing data look like? What rules should it follow? Should there be added noise (unexpected ups and downs) that match the surrounding data or should it be a smoother transition that might be described by a moving average or similar approach? Should the season be a predictor? Could you scrape the data from weather history websites? Should the weather from the preceding n-days be a variable? Once you've defined that, you can determine whether those function are suitable by reading the documentation or by creating your own custom function.

[Walter Roberson]

No-one knows what the best method for estimating missing data is. There is reason to expect that the best method cannot be known until infinite computation has taken place to compare all possible methods under every circumstance.