This document summarizes how a Kohonen network can be used to visualize high-dimensional data in a lower dimensional space while preserving topological similarities. It explains that a Kohonen network consists of input nodes equal to the data dimension and a grid of classification nodes with weight vectors. The learning algorithm initializes node weights, selects a random input vector, finds the best matching unit, adjusts the weights of neighboring nodes based on their distance from the BMU, and repeats for multiple iterations to map similar data points close together on the grid. It then briefly describes how Data-Applied's web interface can be used to select data and perform similarity analysis using a Kohonen network.

1. Data-Applied.com: Similarity

2. IntroductionVisualization of data having high dimensions is very difficultThe kohonen network enables us to represent high dimensional data in lower dimensions while keeping the topological information in it like similarity etc. intactKohonen networks or self organizing maps learn to classify data without supervision unlike artificial neural networksA SOM does not need a target output to be specified unlike many other types of network. The learning steps are:

3. Components of Kohonen networkKohonen networks consist of:Input nodes which are equal in number to the dimension of the dataA grid of nodes which are used for classification, each having a weight vector equal to the dimension of input and connected to all inputs

4. Learn Algorithm OverviewEach node's weights are initialized.A vector is chosen at random from the set of training data and presented to the lattice.Every node is examined to calculate which one's weights are most like the input vector. The winning node is commonly known as the Best Matching Unit (BMU).The radius of the neighbourhood of the BMU is now calculated. This is a value that starts large, typically set to the 'radius' of the lattice, but diminishes each time-step. Any nodes found within this

5. Learn Algorithm OverviewEach node's weights are initialized.A vector is chosen at random from the set of training data and presented to the lattice.Every node is examined to calculate which one's weights are most like the input vector. The winning node is commonly known as the Best Matching Unit (BMU).The radius of the neighbourhood of the BMU is now calculated. This is a value that starts large, typically set to the 'radius' of the lattice, but diminishes each time-step. Any node found within

6. Learn Algorithm Overview(contd..) radius are deemed to be inside the BMU's neighbourhood.5. Each neighbouring node's (the nodes found in step 4) weights are adjusted to make them more like the input vector. The closer a node is to the BMU, the more its weights get altered.6. Repeat step 2 for N iterations.

7. Similarity using Data Applied’s web interface

8. Step1: Selection of data

9. Step2: Selecting Similarity

10. Step3: Result

11. Visit more self help tutorialsPick a tutorial of your choice and browse through it at your own pace.

12. The tutorials section is free, self-guiding and will not involve any additional support.

13. Visit us at www.dataminingtools.net