I have the data set of daily price of 8 fuels types like as Heating Oil, Brent, Methanol and so on from 7.11.2002 to 7.11.2022.
Now I would like to predict their day prices for the coming days. For example, for Brent fuel, I want to apply ANFIS for prediction of daily price. I have the daily price of the 20 years in the Excel file (I sent its image in the attachment below). I want to predict the daily price of this fuel for the upcoming 60 days after the last day (7.11.2022). I considered the $80%$ of data as training data and the rest of them as the testing data. Moreover, I considered one input data which is a matrix with two columns. First column shows the number of index of the day's number and the second column shows the value of daily price.
When I use the "Generate FIS", I encounter the Error: "Xout must be finite".
How can I solve this Error? Should I make change in the structure of my input data?
Thanks a lot.
As you can see in the data, one data point is infinite, and thus, the error message popped up. Identify the location and remove it should get the job done.
x = [1 2 3 4 5];
y = [2 4 6/0 8 10];
data = [x' y']
[row, col] = find(isinf(data))
If one or more of the array elements are NaN (such as the indeterminate 0/0), then use:
[row, col] = find(isnan(data))
Zeinab Hassanzadeh I encountered the same problem recently. The error occurred because some points in your data are infinite (- Inf or + Inf). You have two options, either fill in these points considering them as missing values, or delete the rows corresponding to these points. start by using the "isinf" function to detect the index of these values (check link bellow).
https://www.mathworks.com/help/matlab/ref/isinf.html
Delete "Inf" or "NaN" values in your input data by hand before uploading the file into Nero-Fuzzy Designer.
Thank you very much. I solved it.
Thanks again for your time, Prof. Dr. Yew-Chung Chak, Berghout Tarek and Eugene Marchuk.
- Badges
- Science topic
- Similar topics
- Engineering
- Control Theory
- Control Systems
- Control Systems Engineering