I have a simulation model of a system which receives a forecast of a time series as input. In my scientific work I would like to examine how the performance of the simulation model behaves in relation to the accuracy of the forecast.

Therefore, I would like to adjust the forecast:

forecast_adjusted = forecast_original.copy()
for i in range(len(forecast_adjusted)):
    forecast_adjusted[i] = forecast_adjusted[i] + p * (actual_data[i] - forecast_adjusted[i])

Where p is a value selected based on the desired improvement or reduction of accuracy.

This can be used to create a time series like the red time series in the following image: enter image description here

My question now is: is this a legitimate way to scientifically examine a system with different forecasts? Or do I lose essential characteristics of the forecast?

  • $\begingroup$ What is the goal here? Why are you adjusting the forecasts? $\endgroup$ – user2974951 Oct 2 at 11:13
  • $\begingroup$ @user2974951 For a system, a operating plan is to be created for the next day. As input into the system forecasts of stock exchanges and weather data are used, which can be forecast according to the state of research with 95% accuracy. However, most users (including me) will only achieve worse predictions, so I would like to analyze how much the prediction accuracy has an effect on the system. $\endgroup$ – Anne Bierhoff Oct 2 at 11:48
  • $\begingroup$ I don't understand. You have a model which makes forecasts based on some data and has 95 % accuracy (whatever that means). Then some users come along and make their own forecasts? And they obtain worse results than the model? And you would like to analyze what? Effect on the system? What does that mean? $\endgroup$ – user2974951 Oct 2 at 12:57
  • $\begingroup$ @user2974951 There are papers where people reach a 95% prediction accuracy (WAPE) on the stock exchange data. Other people like me only achieve about 82% prediction accuracy. The system prefers to work at high stock exchange prices. Now I want to analyze for the system how well it works, depending on how good the forecast for the stock exchange data is. For example, to analyze whether it is worth investing a lot of effort in a good forecast. $\endgroup$ – Anne Bierhoff Oct 2 at 13:24

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.