How to simulate missing data?

Question

Mathematics

There is a large array of data, namely the data taken from the sensors. In the database, some data cells are empty. What are the methods of supplementing the missing data based on the existing ones? Maybe there are ready-made solutions in Excel?
PS average values ​​based on existing data are not suitable because all this is eventually driven into a neuron, and it immediately sees something like this and gives out "they say it's not good to take average values"

Answer 1

Found the solution myself. Who cares. Made in excel. Take the longest contiguous sequence to an empty data cell, calculate the value of the PREDICT function for this cell, then do the same for the shorter sequence and take the average value between the received data and the data of the adjacent cells on both sides. In this case, the value is closer to reality.

Answer 2

Through the inverse function of the degree N (in Excel it is, in the graphs)
pasteboard.co/2ugHAtXZ.png