I want to evaluate the robustness of the clustering algorithm to noise ... How can I add noise to the data... Is there a well known method (such as salt and pepper in the image data)?
you could perturb the values (i.e. flip categories). note that this will be 'heavy' noise.
You can use VICUS, a noise addition technique for categorical data.
the paper can be found here:
https://crpit.scem.westernsydney.edu.au/confpapers/CRPITV134Giggins.pdf
Dear Amin Golzari Oskouei,
I would suggest you do the following:
Assuming you have a total of 100 data samples named "Dataset":
The random noise can be added as follows:
1. compute the random noise and assign it to a variable "Noise"
2. Add the noise to the dataset ( Dataset = Dataset + Noise)
3. Partition the Noisy Dataset into three parts:
a). 70% for training (Training data)
b). 15% for validation (Validation data)
c). 15% for testing (Testing data)
Alternatively, you can partition the Noisy Dataset into two-part:
a). 75% for training (Training data)
b). 25% for testing (Testing data)
4. Then, you can then use a classifier ( Neural Network, SVM, LDA, ...) to classifier the Dataset
5. Finally, You can now examine the classification accuracy of the classifier.
I hope this information will be helpful for you
Dear Amin
Ok in clustering DBSCAN is widely used to detect outliers and noise in your data so assuming that you already have noise in your original data you can use the algorithm to detect it first and then make Fata augmentation using the values obtained as outliers
- Badges
- Science topic
- Similar topics
- Chemistry
- Sulfites
- Inorganic Chemistry
- Inorganic Chemicals
- Salts