Skip to main content

A utility to extract world models of information sources.

Project description

# metaparse

This is to be a machine learning project to extract metainformation about information sources, such as agents online that share information.

## Problem specification

Given some data from an information source I, recognize its [intent about content](https://wefindx.net) towards the world, answering following questions about the information source and its information content:

  • F(domain).assets:

  • .Agent (who?)

  • .Place (where?)

  • .Event (when?)

  • .Topic (about what?)

  • X(process).actions:

  • . doing (doing what?)

  • Y(range).targets:

  • .Goal (why?)

  • .Idea (how in principle?)

  • .Plan (what specifically?)

  • .Step (in what order?)

This will allow to treat every information source as an instance of equation optimization process F(X)=Y, where the actions of agents are explained by their intents (Y) about content their world model (F).

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

metaparse-0.0.1.tar.gz (1.6 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page