linda-0.1.2: LINear Discriminant Analysis

Portability portable experimental lennart...schmitt@gmail.com

Numeric.Statistics.LDA

Description

This module implements some linear discriminant analysis functions. Imagine you've made a poll and now you have values/attributes from every subscriber. Further more you've grouped the subscribers into clusters. The poll-datas are structured as follows:

• poll-data of one subscriber = [value] --> Vector value
• poll-data of one cluster/group of subscribers = [[values]] --> Matrix values
• poll-data of all clusters/groups = [[[values]]] --> MatrixList values

Now you want to check if you clustered right and/or how significant the values you asked for are...

Synopsis

# Documentation

Calculation of the classification of a survey (or attributes) in a cluster. The function takes a vectorlist of attributesvalues and a context. The context consists of groupsclusters and its items valuesattributes. The function returns the ID (starting with 0) of the cluster to which the given vector/list belongs to. This function uses the Fisher algorithm.

Calculates the ID (starting with 0) of the cluster the given list of attributes belongs to. The function takes a list of attributes and a list of clusters which are representated by there classification function. This function uses the Fisher algorithm.

Calculates the ID of the cluster the given values belonging to. This function takes a list of clusters, representated by a tuple, and a list of values. The cluster-tuples consists of a ID of the cluster and the classification function (according to Fisher) of the cluster. This function uses the Fisher algorithm.

Calculates the cluster of every survey of a poll. This function takes the data of a whole poll and classifies every survey of the poll. This function uses the Fisher algorithm.

Calculates the classification function according to Fisher.

Calculation of the a priori probability, more precisely the probability that an element belongs to a group.

Calculates the discriminant criteria.

Calculates the isolated discriminants of every attribute.

``` isolatedDiscriminant [[[-1,1],[2,2]],[[1,3],[4,8]]] == [0.4444444444444444,1.2307692307692308]
```