- Draw a line connecting two centroids and use the half way line as a division line for two hyperplanes (if two clusters). Results vary greatly.
- Initial positions of centroid can strongly influence result. Different initial positions give completely different results.
- Analogy "Rubber Band"
- Center of the cluster is called a centroid
- Number of centroids at initiation can heavily influence the result.
- Great for ... PROS:
- Bad for ... CONS ... limitations:
- Hill climbing algorithm.
- Result depends on initiation
- If initiation is close to local optima, may be sticky. Never move away. Ignore global optima. Bad initial centroids exist
- If there are more potential clusters, there are more local optima. Run iterate the algorithm many times to avoid being stuck.
Saturday, March 25, 2017
K Means Clustering Unsupervised Learning - Udacity Machine Learning Nanodegree Flash Card
There are three key k means clustering parameters in sklearn that you will need to pay attention to: Number of centroids, aka center of c...
What is a domain name system (DNS)? How stuff works explains it in a very good graph I was very confused by the Wikipedia explanatio...
The bogus request from P2PU to hunt for HTML tags in real life has yielded a lot of good thoughts. My first impression was that this is stup...
Dilys Sun Got a question about web development dev bootcamps? Ask them here or @i_stanford Your question shall be answered by myself, othe...