newLISP Development release 10.5.2

[Lutz]

This development release fixes bugs and introduces functions for Kmeans cluster analysis.



With the already existing bayes-train, bayes-query and the new kmeans-train and kmeans-query functions, newLISP now implements 2 of the most - or even the 2 most - used machine-learning algorithms as fast native functions.



Changes notes and files: http://www.newlisp.org/downloads/development/">http://www.newlisp.org/downloads/development/



Ps: Users of 10.5.0 who use the bigint functions system should also switch to v.10.5.2 (the fixes where already in v.10.5.1)

[HPW]

Typo in release notes:

After Jave update 7u21, install directory names could not have spaces.

After Java update 7u21, install directory names could not have spaces.

[Lutz]

Thanks HPW, I wonder if anybody is planning to use the new kmeans-train and kmeans-query functions?

[johu]

Hello, Lutz.



For the manual of Maintenance Release.



line 6548

<td>calculate distances to cluster centroids or other data points</td>

　　　↓

<td>calculates distances to cluster centroids or other data points</td>

line 6553

<td>partion a data set into clusters</td>

　　　↓

<td>partitions a data set into clusters</td>

And

crit-t   calculates the Student's t statistic for a given probability

kmeans-query   calculate distances to cluster centroids or other data points

kmeans-train   partion a data set into clusters

crit-z   calculates the normal distributed Z for a given probability

is

crit-t   calculates the Student's t statistic for a given probability

crit-z   calculates the normal distributed Z for a given probability

kmeans-query   calculate distances to cluster centroids or other data points

kmeans-train   partion a data set into clusters

maybe.



line 16387

syntax: (kmeans-query list-data matix-data</h4>

　　　　　　　　　　　　　　　　　　　　　　　　　　　　　　↓

syntax: (kmeans-query list-data matix-data)</h4>

I think that "Eucledian distance" equals "Euclidean distance".

Both seem to be used in the web, and in Wikipadia

http://en.wikipedia.org/wiki/Euclidean_distance">http://en.wikipedia.org/wiki/Euclidean_distance



I don't know which common.

[Lutz]

Thanks Johu, all corrections online:



http://www.newlisp.org/downloads/development/newlisp_manual.html">http://www.newlisp.org/downloads/develo ... anual.html">http://www.newlisp.org/downloads/development/newlisp_manual.html