Novel Lexicon Hierarchy Semantic Embedding for Domain Specific Text Mining

PDF Version Also Available for Download.

Description

This paper proposes a lexicon hierarchy semantic embedding model for domain text mining.

Physical Description

8 p.

Creation Information

Chen, Xiaoli; Han, Tao & Zhang, Zhixiong November 9, 2018.

Context

This paper is part of the collection entitled: International Conference on Knowledge Management (ICKM) and was provided by UNT College of Information to Digital Library, a digital repository hosted by the UNT Libraries. More information about this paper can be viewed below.

Who

People and organizations associated with either the creation of this paper or its content.

Authors

Provided By

UNT College of Information

Situated at the intersection of people, technology, and information, the College of Information's faculty, staff and students invest in innovative research, collaborative partnerships, and student-centered education to serve a global information society. The college offers programs of study in information science, learning technologies, and linguistics.

Contact Us

What

Descriptive information to help identify this paper. Follow the links below to find similar items on the Digital Library.

Description

This paper proposes a lexicon hierarchy semantic embedding model for domain text mining.

Physical Description

8 p.

Notes

Abstract: Feature representation plays a very important role in text mining tasks, especially for domain-specific scientific documents mining. In this paper, we propose a lexicon hierarchy semantic embedding model (LHSE) for domain text mining. The novel embedding model associates each lexicon pair in the lexicon hierarchy with a distance metric. Lexicon pair’s distance are calculated with respecting to their relative hierarchy in the lexicon hierarchy. We use a tree-based path weight to calculate each lexicon pair’s semantic distance. To test our novel embedding model, we compare our lexicon hierarchy semantic embedding (LHSE) based CNN classification algorithm with other state of the art text classification algorithms. Text classification experiments conducted in chemical domain scientific documents show the superiority of our proposed method.

Source

  • 14th International Conference on Knowledge Management, November 9-10, 2018. Vancouver, Canada.

Language

Item Type

Relationships

Collections

This paper is part of the following collection of related materials.

International Conference on Knowledge Management (ICKM)

Serving as digital proceedings, this collection includes papers, posters, and slides from invited talks as well as practitioner and sponsor presentations for the annual International Conference on Knowledge Management (ICKM).

Related Items

Novel Lexicon Hierarchy Semantic Embedding for Domain Specific Text Mining (Presentation)

Novel Lexicon Hierarchy Semantic Embedding for Domain Specific Text Mining

Presentation for the 2018 International Conference on Knowledge Management. This presentation proposes a lexicon hierarchy semantic embedding model for domain text mining.

Novel Lexicon Hierarchy Semantic Embedding for Domain Specific Text Mining - ark:/67531/metadc1438979

What responsibilities do I have when using this paper?

When

Dates and time periods associated with this paper.

Creation Date

  • November 9, 2018

Added to The UNT Digital Library

  • Dec. 19, 2018, 12:07 p.m.

Usage Statistics

When was this paper last used?

Yesterday: 1
Past 30 days: 1
Total Uses: 7

Interact With This Paper

Here are some suggestions for what to do next.

Start Reading

PDF Version Also Available for Download.

International Image Interoperability Framework

IIF Logo

We support the IIIF Presentation API

Chen, Xiaoli; Han, Tao & Zhang, Zhixiong. Novel Lexicon Hierarchy Semantic Embedding for Domain Specific Text Mining, paper, November 9, 2018; (https://digital.library.unt.edu/ark:/67531/metadc1393782/: accessed October 17, 2019), University of North Texas Libraries, Digital Library, https://digital.library.unt.edu; crediting UNT College of Information.