Spectrep: A software for spectral representation of text content

Abstract

Spectrep (Spectral Representation) implements a text representation framework for obtaining a multi-view representation that consolidates lexical, syntactic, and semantic features by structuring a workflow with popular text processing tools such as NTLK, Gensim, and Self-Organizing Maps. This framework is intended to be used in text pre-processing for its application in more robust machine learning applications. This software is open-source and implemented in Python. It has been successfully tested with several tasks to prove its functionality.

Keywords

• Multi-view representation
• Machine Learning
• Content spectra