Skip to content

Latest commit

 

History

History
25 lines (21 loc) · 642 Bytes

TODO.md

File metadata and controls

25 lines (21 loc) · 642 Bytes

Ideas

Data structure

Each record contains at least one label.

{
  "content": "A piece of text that needs to be processed",
  "labels": {
    "category": "business/sport/politics etc."
    "emotions": "positive/negative/neutral"
    "language": "official/urban"
  }
}

Each record contains at least one label.

Predicting process

Many models, each one recognizes one label (category, emotions, language). Models are automatically learned and added to "Prediction chain" when new label is found.

Presentation

Chrome extension. When selecting any text it predict its labels.

Clustering

Data from https://newsapi.org/