mljs / kmeans Goto Github PK

• kmeans++
• Random partition

Hi,

I'm using Rollup to build an es6 module and when I go to build my files, I get the following error:

[!] Error: 'squared' is not exported by node_modules/ml-kmeans/node_modules/ml-distance-euclidean/euclidean.js
npm ERR! code ELIFECYCLEp/rollup/wiki/Troubleshooting#name-is-not-exported-by-module
npm ERR! errno 1kmeans/src/initialization.js (2:9)

I realize this could probably be handled using rollup-plugin-commonjs, but it looks like the issue has already been partially addressed in version 4.1.

Your distance-euclidean repo builds a directory lib-es6 but this is not present in node_modules/ml-kmeans/node_modules/ml-distance-euclidean/ so it ends up using the default commonjs version.

Am I missing something here? Has ml-kmeans been tested with an es6 build?

Thanks,
Josh

I'd like to refactor the library to a class, with an API similar to sklearn: http://scikit-learn.org/stable/modules/generated/sklearn.cluster.KMeans.html

I'll migrate to TypeScript at the same time.

@mljs/collaborators WDYT?

function kmeans(data, k, options)

options.initialisation

• array of your current centers
• 'random' (default value) : we take randomly k points of the data

• weighted kmeans
• triangle inequality
• kernel kmeans
• kmedian

I tried to use ml-kmeans in a TypeScript project, but it appears that kmeans I import is not a function, generating .d.ts file may help.

In the userguide it seems it works on with x,y points. Is this correct ?

I would like to kluster r,g,b points.

First of all, thanks for the amazing library.

1. How cumbersome would it be to add the wcss to the KMeansResult?
2. Would it be possible to implement a predict method?

Thanks again!

Error message that states, "K should be a positive integer bigger than the number of points" should say "smaller", rather than "bigger". Line 76 src/kmeans.js

Happy to fix this and submit a pull request, if preferred. Didn't think an error message typo may be worthy of a PR. :-)

• Convert code to typescript
• Generate tsdoc

Sometimes KMeansResult's centroids value is NaN for both x,y (Two-dimensional array) and for this centroid error value is NaN.