Pandas how to do date continuation for forecast?

Question

Python

Hello, I wrote a neuron for predicting time intervals. And something stuck on the visualization of the forecast. Let's admit there is some value and date when it is received. Let's say I take the last 200 values ​​from the array, the neuron predicts the next 200, how to correctly continue the date for each number without complex calculations by type, the difference between the value, how many values ​​per day, per week and how the time difference, and so on, to then supplement the predicted values ​​with a new one X-axis time

Answer 1

First, are you sure you are "forecasting time INTERVALS"? In fact, the forecasting of exactly the intervals, i.e. time BETWEEN two events are performed using the so-called. series of events, which, as a rule, have a Poisson distribution. Are you sure this is what you do?
Secondly, try to read this for yourself: " how to correctly continue the date for each number without complex calculations by type the difference between the value, how many values ​​per day, per week and how the time difference, and so on, so that later the predicted values ​​\u200b\u200bare supplemented with a new time according to X-axis " And try to understand what is written here.
Thirdly, as far as I understand from your many previous posts on this and other forums, you actually have typical equidistant rows. And in fact, you predict not from the date, but from the tick number. And if so, then "continue the series for forecasting" is to take ticks, starting from the last one for which you trained the network and further, with a given interval, for how long the "arrogance" is enough. The latter is not a joke, one should always remember about the permissible forecasting horizon and the extreme undesirability of going beyond it.