On improving imputation accuracy of LTE spectrum measurements data

Univariate imputation, such as Kalman filtering, is not able to provide a reasonable imputation for a variable when periods of missing values are large. A new method is needed that can provide feasible imputations in such scenarios. We propose a novel method of applying multivariate imputation in co...

Full description

Saved in:
Bibliographic Details
Published in2018 Wireless Telecommunications Symposium (WTS) pp. 1 - 7
Main Authors Chaudhry, Aizaz, Li, Wei, Basri, Amir, Patenaude, Francois
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.04.2018
Subjects
accuracy
Frequency measurement
imputation
Kalman filters
Long Term Evolution
LTE
multivariate
Object recognition
Throughput
univariate
Online AccessGet full text

Cover

More Information
Summary:Univariate imputation, such as Kalman filtering, is not able to provide a reasonable imputation for a variable when periods of missing values are large. A new method is needed that can provide feasible imputations in such scenarios. We propose a novel method of applying multivariate imputation in combination with an existing univariate imputation approach to a single variable in an LTE spectrum dataset, such as the average cell throughput, by exploiting the high weekly seasonality of this variable. Performance comparison shows that our proposed method significantly outperforms Kalman filtering in terms of imputation accuracy.
DOI:10.1109/WTS.2018.8363929