I've been working on integrating R into Qlik for a long time and using the Advanced Analytics Toolbox visualization extension. I am using it essentially for forecasting but currently I have come across a case where I cannot find an explanation for the prediction provided by the extension.
I am going to attach a screenshot below. I do not understand the prediction it makes (clearly underestimated compared to the real data). It seems very strange to me and I don't know why. In what can be seen in the image, no filter has been applied and the real data of the measurement is well loaded (verified). It is a monthly series since 2000. The top chart is Arima (parameter settings in auto option) and the one below is Holt Winters.
I would like to understand the problem and fix it. Thank you very much for your attention.