Luiz Andrade

In this paper (https://www.researchgate.net/publication/283185453_Analysis_of_KDD_Dataset_Attributes_-_Class_wise_for_Intrusion_Detection) you can find the features used in the dataset.

Hi there! I think the default behavior should be normalize to [0, 1] according to the data type used. This would avoid confusion. I also think this information should be...

I've made a CustomGeoDatamodule like this: ``` class CustomGeoDataModule(GeoDataModule): def setup(self, stage: str) -> None: """Set up datasets. Args: stage: Either 'fit', 'validate', 'test', or 'predict'. """ self.dataset = self.dataset_class(**self.kwargs)...

Using `min_tsep=1` it works.... But I would need a good approach for min_tsep, that would be the mean period, but how to make it work?

Thanks for your help! Let me give one of the datasets, the respective yaml and the code I'm using. This is one of the datasets: [el_carmen.csv](https://github.com/user-attachments/files/17209158/el_carmen.csv) This is the yaml...

Would you agree with this? ``` mtsep = np.ceil(1/mf.real) mtsep = min(mtsep, int(0.25 * n)) ```

Thanks for your help!!! Changing to `poly` gave me no changes. The debug plots seem ok to me. Here some examples: R/S: [el_carmen_hurst_rs_0.pdf](https://github.com/user-attachments/files/17210014/el_carmen_hurst_rs_0.pdf) [el_carmen_hurst_rs_1.pdf](https://github.com/user-attachments/files/17210016/el_carmen_hurst_rs_1.pdf) DFA: [el_carmen_hurst_dfa_0.pdf](https://github.com/user-attachments/files/17210020/el_carmen_hurst_dfa_0.pdf) [el_carmen_hurst_dfa_1.pdf](https://github.com/user-attachments/files/17210021/el_carmen_hurst_dfa_1.pdf) GHE: [el_carmen_hurst_ghe_0.pdf](https://github.com/user-attachments/files/17210025/el_carmen_hurst_ghe_0.pdf) [el_carmen_hurst_ghe_1.pdf](https://github.com/user-attachments/files/17210026/el_carmen_hurst_ghe_1.pdf)

One question, should I remove the mean from the series before running Hurst?

An improvement. I've changed the nvals calculation to the classical division by powers of 2. I was using your `logmid_n` method as follows: ``` nstepss = np.arange(15, 31) nvalss =...