Skip to content

Latest commit

 

History

History
86 lines (58 loc) · 2.55 KB

README.md

File metadata and controls

86 lines (58 loc) · 2.55 KB
Raw

glowfi.sh in Android for the Big Web: Now with machine guns and rocket launchers.

Installation

Download glowfish.aar and drop it in your app...seriously, that's it.

Setup

GFGlower.glower().setCredentials('<GLOWFISH_SID>', '<GLOWFISH_AUTH_TOKEN>');

Useage

Get ready for some simple machine learning...

Training

GFGlower.glower().train(new JSONObject("{
    "feature_name1": [1, 2, 3, 4, ...etc],
    "feature_name2": [9, 4, 5, 6, ...etc]
}"), new JSONObject("{
    "class": [4, 3, 5, 6, ...etc]
}"), new JSONObject("{"param1": "value"}"), <GFResponse>);

Predict It's important to note that predicting will throw an error if you have not trained against a data set first.

GFGlower.glower().train(new JSONObject("{
    "feature_name1": [1, 2, 3, 4, ...etc],
    "feature_name2": [9, 4, 5, 6, ...etc]
}"), new JSONObject("{
    "class": [4, 3, 5, 6, ...etc]
}"), new JSONObject("{"param1": "value"}"), <GFResponse>);

Clustering

GFGlower.glower().train(new JSONObject("{
    "feature_name1": [1, 2, 3, 4, ...etc],
    "feature_name2": [9, 4, 5, 6, ...etc]
}"), new JSONObject("{"param1": "value"}"), <GFResponse>);

Feature Selection

GFGlower.glower().train(new JSONObject("{
    "feature_name1": [1, 2, 3, 4, ...etc],
    "feature_name2": [9, 4, 5, 6, ...etc]
}"), new JSONObject("{
    "class": [4, 3, 5, 6, ...etc]
}", new JSONObject("{"param1": "value"}"), <GFResponse>);

Filter Train

GFGlower.glower().train(new JSONArray("[1, 2, 3, 4, ...etc]"), new JSONArray("[4, 3, 5, 6, ...etc]", new JSONArray("[4, 3, 5, 6, ...etc]", new JSONObject("{"param1": "value"}"), <GFResponse>);

Filter Predict

GFGlower.glower().train(new JSONArray("[1, 2, 3, 4, ...etc]"), new JSONArray("[4, 3, 5, 6, ...etc]", new JSONArray("[4, 3, 5, 6, ...etc]", new JSONObject("{"param1": "value"}"), <GFResponse>);

Responses

The final parameter of all calls should be a class that implements GFResponse. A GFResponse interface containts only one method:

@Override
public void response(GFGlower.GFApiResponse response) {
  // response is the object to use
}

The GFApiResponse class is a simple map class:

public class GFApiResponse {
    JSONObject results;
    JSONObject metrics;
    JSONObject errors;
    private JSONObject statusObj;

    Integer status;
}

Further Documentation

Docs - http://glowfish.readme.io/
Registration - http://glowfi.sh/

Thank You

Thanks so much to Matt Thompson (@matt) for creating Alamofire. Big props.