Computer scienceData scienceMachine learningClustering

DBSCAN

Parameter initialization

Report a typo

What are the factors to consider when selecting the minimum points and the $\epsilon$ parameters?

Hint

Carefully review the Hyperparameter selection section of the theory.

Select one or more options from the list

The presence of noise (when there are many outliers, the minimum points parameter should be increased)

The minimum points parameter is usually set to the number of samples in the dataset

When the dataset is very large, high dimensional, or contains many duplicates, the minimum points parameter could be increased

The data density (e.g., when the data is sparse, the epsilon should be increased, and vice versa for a dense dataset)

In general, the minimum points parameter is set to 2*dim

The minimum points parameter could be set at random

The k-distance plot might be useful in determining the optimal epsilon value

___

Create a free account to access the full topic