Link quality estimation method based on gradient boosting decision tree
by Yan Zhang; Jian Shu
International Journal of Sensor Networks (IJSNET), Vol. 36, No. 3, 2021

Abstract: In the application of wireless sensor networks, link quality estimation is the primary problem to guarantee the reliable transmission of data and the performance of the upper layer network protocol. In order to accurately evaluate link quality, a link quality estimator based on gradient boosting decision tree (GBDT) was proposed. The physical layer parameter average received signal strength indication, mean link quality indicator and mean signal noise rate is selected as the input of the GBDT estimator and the nonlinear correlation between physical layer parameters and packet received rate is analysed by using the maximum information coefficient method. Considering the influence between outliers and different dimensionality of parameters, we used the boxplot method to carry out smoothing and normalisation processing to reduce the complexity of the estimator. At last, the improved particle swarm optimisation algorithm is used to select the optimal parameter combination in the GBDT estimator. The experimental results show that compared with the support vector machine (SVM) estimator, the estimator of this paper has higher accuracy and stability.

Online publication date: Tue, 24-Aug-2021

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 Sensor Networks (IJSNET):
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