Premium
Tweedie regression model: a proposed statistical approach for modelling indoor signal path loss
Author(s) -
Hassine Aymen,
Masmoudi Afif,
Ghribi Abdelaziz
Publication year - 2017
Publication title -
international journal of numerical modelling: electronic networks, devices and fields
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.249
H-Index - 30
eISSN - 1099-1204
pISSN - 0894-3370
DOI - 10.1002/jnm.2243
Subject(s) - regression analysis , statistics , mathematics , linear regression , variance (accounting) , generalized linear model , regression , accounting , business
In this paper, we focused on the modelling of the signal path loss prediction by using the Tweedie regression model. The distance and the obstacles quantity between receiver and transmitter are chosen as explanatory regression variables. This enables us to compare it with the log‐normal shadowing model. The Tweedie regression model is a wide case of the generalized linear model with a Tweedie response distribution characterized by the unit variance function with power mean‐variance relationships μ p , p ∈ R ∖ ( 0 , 1 ) . We studied the asymptotic normality and the confidence interval of the predicted signal path loss. The performance of this model was studied using a real data sample to select the fitted regression model by varying the Tweedie parameter p . We illustrated our approach by considering an experimental study and using the bootstrap technique to select the appropriate Tweedie regression model. The experimental results showed that the proposed approach outperforms the prediction of signal path loss.