- To undo the outlier elimination, you can use the 'rmoutliers' function in MATLAB, which removes the outliers from the data. You can use the same method you used for outlier elimination (i.e., 'pchip' and 'grubbs') to remove the outliers from your simulated values.
- To undo the smoothing, you can use the smoothdata function in MATLAB with the 'movmedian' option as you did earlier, but this time use the 'rloess' option instead. This option applies a robust local regression method that can undo the smoothing performed by the 'movmedian' option.
How to obtain the simulated and unpreprocessed version of the data?
1 次查看(过去 30 天)
显示 更早的评论
I am performing the following preprocessing steps. First, I elliminate the outliers with filloutliers(data, 'pchip', 'grubbs'). Then, I smooth the data with smoothdata(data, 'movmedian'). I also perform scaling and db2mag conversion. After training the feedforwardnet with [8 8 8] number of neurons, I obtain the simulated values of training and testing sets. I now need to obtain the simulated values of these data in the before the preprocessing state. In terms of scaling and db2mag conversion it is straightforward but in case of filling the outliers and smoothing I am confused. How can I solve this problem?
0 个评论
回答(1 个)
Rijuta
2023-3-24
Hi Miras,
I understand that you want to obtain simulated values of your data in the pre-processed state.
For scaling and db2mag conversion, this is straightforward as you mentioned. However, for outlier elimination and smoothing, you will need to apply the inverse transformations. In other words, you will need to undo the outlier elimination and smoothing that you performed on your data.
Once you have undone the outlier elimination and smoothing, you should have simulated values in the pre-processed state. I hope this solves your query.
0 个评论
另请参阅
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!