ARIMA Model Forecasting issue

I am trying to creat an ARIMA model to predict future data points. Using the Econometrics App I fitted a model ARIMA(4,1,3) that fits the data very well. I then try to forcast the data using forcast.

SteelData1Log = log(SteelData1);

SteelData1LogDiff = [NaN; diff(SteelData1Log)];

ARIMA_SteelData1LogDiff2 = arima('Constant',NaN,'ARLags',1:4,'D',1,'MALags',1:3,'Distribution','Gaussian');

ARMA = estimate(ARIMA_SteelData1LogDiff2,SteelData1LogDiff,'Display','off');

yF1 = forecast(ARMA,44,'Y0',August2017.Var2(1:100));

The result does not follow the data, and has a logorithmic trend to it.

[BN]

Dear @Michael MuellerMichael,

Did you find an answer for this problem?

[Stephen Gray]

I too am looking for an answer to this!