feature-mining 0.1.1

Mine implicit features using a generative feature language model.

GFLM: mine implicit features using a generative feature language model

Description

This package implements a Generative Feature Language Models for Mining Implicit Features.

Given the following input:

• a text dataset
• a set of predefined features

Compute the following:

• mapping of explicit and implicit features on the data
• using both gflm_word and gflm_section algorithms

Install

pip install feature_mining

Sample Usage

Usage:
    from feature_mining import FeatureMining
    fm = FeatureMining()
    fm.load_ipod(full_set=False)
    fm.fit()
    fm.predict()

Results:
    - prediction using 'section': fm.gflm.gflm_section
    - prediction using 'word': fm.gflm.gflm_word

Display result:
    fm.section_features()
    print(fm.gflm_section_result.sort_values(by=['gflm_section'], ascending=False)[['feature', 'section_text']].head(20))

Package created based on the following paper

S. Karmaker Santu, P. Sondhi and C. Zhai, "Generative Feature Language Models for Mining Implicit Features from Customer Reviews", Proceedings of the 25th ACM International on Conference on Information and Knowledge Management - CIKM '16, 2016.

Pydocs (Code Documentation)

Accessible via this link: http://htmlpreview.github.io/?https://github.com/nfreundlich/CS410_CourseProject/blob/dev/docs/feature_mining.html

(Apologies for the color scheme - it was the default)

Tutorial

See Jupyter notebook tutorial https://github.com/nfreundlich/CS410_CourseProject/blob/dev/tutorial.ipynb

Video presentation and tutorial

Link to YouTube: https://www.youtube.com/watch?v=mjJHkyrkxHM

Package on PyPi

https://pypi.org/project/feature-mining/

Slides

https://github.com/nfreundlich/CS410_CourseProject/blob/dev/docs/CS_410_GFLM_Slides.pdf

Known Issues

Explicit feature mentions not removed from GFLM word/sentence: https://github.com/nfreundlich/CS410_CourseProject/issues/28