Skip to main content

A library for doing incremental concept formation using algorithms in the COBWEB family.

Project description

https://travis-ci.com/cmaclell/concept_formation.svg?branch=master https://coveralls.io/repos/github/cmaclell/concept_formation/badge.svg?branch=master

Concept Formation

This is a Python library of algorithms that perform concept formation written by Christopher MacLellan (http://www.christopia.net) and Erik Harpstead (http://www.erikharpstead.net).

Overview

In this library, the COBWEB and COBWEB/3 algorithms are implemented. These systems accept a stream of instances, which are represented as dictionaries of attributes and values (where values can be nominal for COBWEB and either numeric or nominal for COBWEB/3), and learns a concept hierarchy. The resulting hierarchy can be used for clustering and prediction.

This library also includes TRESTLE, an extension of COBWEB and COBWEB/3 that support structured and relational data objects. This system employs partial matching to rename new objects to align with previous examples, then categorizes these renamed objects.

Lastly, we have extended the COBWEB/3 algorithm to support three key improvements. First, COBWEB/3 now uses an unbiased estimator to calculate the standard deviation of numeric values. This is particularly useful in situations where the number of available data points is low. Second, COBWEB/3 supports online normalization of the continuous values, which is useful in situations where numeric values are on different scales and helps to ensure that numeric values do not impact the model more than nominal values. Finally, it is assumed that there is some base noise in measuring continuous values, this noise insures that the probability of any one value never exceeds 1, even when the standard deviation is small.

Installation

You can install this software using pip:

pip install -U concept_formation

You can install the latest version of the code directly from github:

pip install -U git+https://github.com/cmaclell/concept_formation@master

Examples

We have created a number of examples to demonstrate the basic functionality of this library. The examples can be found here.

Citing this Software

If you use this software in a scientific publiction, then we would appreciate citation of the following paper:

MacLellan, C.J., Harpstead, E., Aleven, V., Koedinger K.R. (2016) TRESTLE: A Model of Concept Formation in Structured Domains. Advances in Cognitive Systems, 4, 131-150.

Bibtex entry:

@article{trestle:2016a,
author={MacLellan, C.J. and Harpstead, E. and Aleven, V. and Koedinger, K.R.},
title={TRESTLE: A Model of Concept Formation in Structured Domains},
journal={Advances in Cognitive Systems},
volume={4},
year={2016}
}

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

concept_formation-0.3.9.tar.gz (408.7 kB view details)

Uploaded Source

Built Distribution

concept_formation-0.3.9-py2.py3-none-any.whl (296.3 kB view details)

Uploaded Python 2 Python 3

File details

Details for the file concept_formation-0.3.9.tar.gz.

File metadata

  • Download URL: concept_formation-0.3.9.tar.gz
  • Upload date:
  • Size: 408.7 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.4.1 importlib_metadata/4.6.1 pkginfo/1.7.1 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.61.2 CPython/3.10.0a5+

File hashes

Hashes for concept_formation-0.3.9.tar.gz
Algorithm Hash digest
SHA256 629e852df2ec455091b0303629ba7cd2ff74efccbbb5d9796bc40e13607e260c
MD5 9ca3634edee1edff8fb4d53269374576
BLAKE2b-256 f47c640329633d00dd0c2c47d5decc66f1b216110d0d8f5a0bef0a906e6a71db

See more details on using hashes here.

File details

Details for the file concept_formation-0.3.9-py2.py3-none-any.whl.

File metadata

  • Download URL: concept_formation-0.3.9-py2.py3-none-any.whl
  • Upload date:
  • Size: 296.3 kB
  • Tags: Python 2, Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.4.1 importlib_metadata/4.6.1 pkginfo/1.7.1 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.61.2 CPython/3.10.0a5+

File hashes

Hashes for concept_formation-0.3.9-py2.py3-none-any.whl
Algorithm Hash digest
SHA256 45113d0d4e4efd3507715927c4473d99c81055ee4b2fc17c038ee51e4a29bbf5
MD5 0ae716316f09555a23ac938a468e51e9
BLAKE2b-256 ad273590cb80e84e784e156cccb025668dd6f7b9bb78eaf1244c49871997736f

See more details on using hashes here.

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