gaspersic/freeCodeCamp
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Updated index.md for Clustering Algorithms (#19099)

Browse Source 
Added description for the types of Hierarchical Clustering and kept it brief and simple.
This commit is contained in:
Tusharkanth Karlapudi
2018-10-15 01:09:10 +00:00
committed by Randell Dawson
parent 8fd54d9d48
commit f5d020a2a2
1 changed files with 6 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
client/src/pages/guide/english/machine-learning/clustering-algorithms/index.md
View File

@ -24,7 +24,7 @@ Some applications of clustering algorithms include:
4. Density-based clustering
Some examples of clustering algorithms are:
1. Alogmerative clustering
1. Agglomerative clustering
2. K-means clustering
3. K-mediods clustering
4. Partition Clustering
@ -34,6 +34,10 @@ There are methods for clustering that only use similarities of instances, withou
This needs the use of a similarity, or equivalently a distance, measure defined between instances. Generally Euclidean distance is used, where one has to make sure that all attributes have the same scale.
There are two main types of Hierarchical clustering which are used:
1. Agglomerative Clustering - This algorithm starts with a bunch of individual clusters and a proximity matrix. Here, the individual clusters are basically individual points, and the matrix is for the distance between each point with each other points. The algorithm tries to find the closest pair of clusters and then combines them into one cluster, and then update the proximity matrix with the new cluster and removes the two combined clusters. This step is repeated until a single cluster is left. The most important part of this algorithm is the proximity matrix and it's updatation.
2. Divisive Clustering - This algorithm can be called an opposite of Agglomerative in terms of how it approachs clustering. It starts with a single cluster and then starts dividing it into multiple clusters. It has a similarity matrix between each point, similarity here being how close the clusters are with each other. This algorithm tries to divide the cluster into two clusters based on how dissimilar a cluster or a point is from the rest. This is continued until there are multiple individual clusters.
### Point Assignment
This method maintains a set of clusters, and it places points to nearest clusters.
@ -120,4 +124,4 @@ Convergence is assured since the algorithm is guaranteed to increase the likelih
* [Cluster Analysis: Basic Concepts and Algorithms](https://www-users.cs.umn.edu/~kumar/dmbook/ch8.pdf)
* [K-means Clustering](https://www.datascience.com/blog/k-means-clustering)
* [Expectation-Maximization Algorithm](https://www.cs.utah.edu/~piyush/teaching/EM_algorithm.pdf)
* [Using K-Means Clustering with Python](https://code.likeagirl.io/finding-dominant-colour-on-an-image-b4e075f98097)
* [Using K-Means Clustering with Python](https://code.likeagirl.io/finding-dominant-colour-on-an-image-b4e075f98097)