Age Prediction in Social Networks Based on Word Embedding and Tensor Learning

DOI

10.2991/ceie-16.2017.28How to use a DOI?

Keywords

Tensor Space Model; Word Embedding; Age Prediction; Tensor Learning; Social Network

Abstract

Latent attribute prediction problem in social network provides a set of conditions for the construction of text classification models. The general framework of current latent attribute prediction problems is mapping the text in social network to vector space, along with a classification model to classify different categories. Unfortunately, as the vector space model ignores the similarity and relevance between different words, it fails to identify the semantic fuzziness in natural language and always performs badly on a long text. With the aim of finding a better framework for age prediction problem, in this paper we propose a word embedding based tensor space model which maps text to tensor feature space. The proposed method relies on supervised tensor learning algorithms which are well studied by many scholars, thus allowing for its easy application in text classification problems. Two experiments on different testing sets show the effectiveness and limitation of our approach.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Ziyi Lin
AU  - Yan Wang
PY  - 2016/10
DA  - 2016/10
TI  - Age Prediction in Social Networks Based on Word Embedding and Tensor Learning
BT  - Proceedings of the International Conference on Communication and Electronic Information Engineering (CEIE 2016)
PB  - Atlantis Press
SP  - 213
EP  - 222
SN  - 2352-5401
UR  - https://doi.org/10.2991/ceie-16.2017.28
DO  - 10.2991/ceie-16.2017.28
ID  - Lin2016/10
ER  -