Understanding the Influence of Hyperparameters on Text Embeddings for Text Classification Tasks

Nils Witt, Christin Seifert

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

7 Citations (Scopus)
17 Downloads (Pure)

Abstract

Many applications in the natural language processing domain require the tuning of machine learning algorithms, which involves adaptation of hyperparameters. We perform experiments by systematically varying hyperparameter settings of text embedding algorithms to obtain insights about the influence and interrelation of hyperparameters on the model performance on a text classification task using text embedding features. For some parameters (e.g., size of the context window) we could not find an influence on the accuracy while others (e.g., dimensionality of the embeddings) strongly influence the results, but have a range where the results are nearly optimal. These insights are beneficial to researchers and practitioners in order to find sensible hyperparameter configurations for research projects based on text embeddings. This reduces the parameter search space and the amount of (manual and automatic) optimization time.
Original languageEnglish
Title of host publicationResearch and Advanced Technology for Digital Libraries
Subtitle of host publication21st International Conference on Theory and Practice of Digital Libraries, TPDL 2017, Thessaloniki, Greece, September 18-21, 2017, Proceedings
EditorsJaap Kamps, Giannis Tsakonas, Yannis Manolopoulos, Lazaros Iliadis, Ioannis Karydis
PublisherSpringer
Pages193-204
ISBN (Electronic)978-3-319-67008-9
ISBN (Print)978-3-319-67007-2
DOIs
Publication statusPublished - 2017
Externally publishedYes
Event21st International Conference on Theory and Practice of Digital Libraries 2017 - Grand Hotel Palace, Thessaloniki, Thessaloniki, Greece
Duration: 18 Sept 201721 Sept 2017
Conference number: 21
http://www.tpdl.eu/tpdl2017/

Publication series

NameLecture Notes in Computer Science
PublisherSpringer
Volume10450
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference21st International Conference on Theory and Practice of Digital Libraries 2017
Abbreviated titleTPDL 2017
Country/TerritoryGreece
CityThessaloniki
Period18/09/1721/09/17
Internet address

Keywords

  • n/a OA procedure

Fingerprint

Dive into the research topics of 'Understanding the Influence of Hyperparameters on Text Embeddings for Text Classification Tasks'. Together they form a unique fingerprint.

Cite this