https://github.com/jaredly/guided-ml

Guided Machine Learning App
https://github.com/jaredly/guided-ml

Last synced: 17 days ago
JSON representation

Guided Machine Learning App

• Host: GitHub
• URL: https://github.com/jaredly/guided-ml
• Owner: jaredly
• Created: 2014-04-02T15:58:30.000Z (over 10 years ago)
• Default Branch: master
• Last Pushed: 2014-05-09T00:04:40.000Z (over 10 years ago)
• Last Synced: 2024-10-13T19:15:07.437Z (about 1 month ago)
• Language: JavaScript
• Homepage:
• Size: 734 KB
• Stars: 1
• Watchers: 3
• Forks: 0
• Open Issues: 1
• Metadata Files:
  • Readme: Readme.md

Awesome Lists containing this project

README

        

# Server

- projects

- features

- learners

- run learners

- get code

```

POST /projects/

- name

- file

=> full data

POST /projects/{id}

- name (or)

- file

=> Ok

POST /projects/{id}/features/

- name

- type

- args

=> new data column

POST /projects/{id}/features/{id}

- name/args

=> changed data column? otherwise 203

DEL /projects/{id}/features/{id}

=> 203

POST /projects/{id}/learners/

- name

- type

- args

=> 203

POST /projects/{id}/learners/{id}

- name or args

=> 203

POST /projects/{id}/train/{id}

=> accuracy, confusion matrix

GET /projects/{id}/compiled/{lang}/{id}

```

# Data Model

```

project: {

  name: '',

  version: 0,

  classes: [],

  headers: [],

  sources: [{

    id: id,

    type: 'segmented',

    filename: '',

    name: '',

  } | {

    id: id,

    type: 'unsegmented',

    filename: '',

    name: '',

    segmenter: {???}

  },

  features: [{

    id: id,

    type: '',

    name: '',

    args: {}

  }],

  processors: [{

    id: id,

    name: '',

    type: '',

    args: ''

  }],

  learners: [{

    id: id,

    name: '',

    type: '',

    args: {}

  }]

},

results: [{

  type: 'single',

  project: {},

  result: {

    accuracies: {},

    confusion: {},

    assignments: {},

    extra: {} // <- defined && used on a plugin level

  }

} | {

  type: 'parameterization', // easily parrallelizable

  object: {

    type: 'processor',

    which: 2,

    param: 'n_big'

  },

  range: {

    from: 1,

    to: 6,

    step: 1

  } | [1,2,3,4,5,6], // <-- for ordinal values too.

  project: {},

  results: {

    // value is one from the range given

    value: {

      accuracies: {},

      confusion: {},

      assignments: {},

      extra: {} // <- defined && used on a plugin level

    }

  }

} | {

  type: 'search', // this can't be parrallelized as well generally.

  project: {},

  object: {

    type: 'feature',

    which: 2,,

    param: 'n_big'

  },

  search_type: ??? what types of searches are there? And could there be a

               interactive search? That's one thing that could be super interesting.

  results: {}

}],

runtime: {

  raw_data: [

    {filename:, meta:, data:, vclass:},

  ],

  data: [[]],

}

```