Sub-word attention mechanism and ensemble learning-based semantic annotation for heterogeneous networks
by Liang Zhang; Zhaobin Liu; Jinxiang Li; Gang Liu; Yuanfeng Yang; Yi Jin; Xu Zhang
International Journal of Wireless and Mobile Computing (IJWMC), Vol. 18, No. 1, 2020

Abstract: The sensing device and wireless sensor networks (WSN) can provide information to the application of Internet of Things (IoT), but data from different types of devices present significant polyphyly and heterogeneity, which poses challenges to the collaboration and interaction of information resources in IoT application and service. Especially when the device uses Chinese characters for information representation and there is an Out-of-Vocabulary (OOV) problem, it will make this work more challenging. This paper introduces an ensemble learning model based on sub-word attention mechanism and bidirectional long short-term memory model (SWAT-Bi-LSTM) which can provide an internal structural attention ability of Chinese characters. The boosting strategy and gradient boosting decision tree (GDBT) integration scheme is adopted to complete the final integrated output. The experimental results show that the proposed method can effectively improve the accuracy of sentiment analysis, and the integrated learning model can further improve the accuracy and stability.

Online publication date: Thu, 30-Jan-2020

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Wireless and Mobile Computing (IJWMC):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com