When develop a model for Intrusion detection system using the deep learning algorithms on determining dataset, it's possible to get the same performance when using another data.?
The performance of Deep Learning based IDS depends on which hyper parameters you are using to optimize the model for the given set of training samples. However, the performance metrics for two different datasets would be different because they have different data distribution.
Following reference give a better understanding for the same:
Chapter Analyses on Influence of Training Data Set to Neural Network...
Unfortunately, it s very hard to prepare one general algorithm that may suitable or leading for every data. For example the deep learning algorithms, are good for large data sets, however., for small datasets it have poor performance.
For optimal algorithms, we first need mathematical/statistical models based on very rich geometrical properties, and good data cleaning etc.
I agree with Ritika Lohiya. The performance of the DL model for two different datasets would be different because they have different data distributions, size and the quality of the data is different.
- Badges
- Science topic