Abstract
Wireless user perception (WeUP) is considered one of the most important factors in designing next-generation wireless communications systems. The recognition of WeUP involves lots of labor cost up to now. In order to solve this problem, this paper proposes an accuracy recognition algorithm for WeUP based on dimensional reduction and semi-supervised clustering. Usually, the WeUP is highly reflected in the key quality indicator (KQI). we build up a database of KQI including more than 1000 cells to train a deep belief autoencoder (DBA) for dimensional reduction (DR). Then we feed the historical unlabeled and manual-labeled negative data set after dimensional reduction into semi-supervised clustering model. After that, we find out a recognition range, which is the most similar to manual-labeled objects with unsatisfied WeUP. Simulation results show that our proposed method achieves an accurate recognition of unsatisfied WeUP over 93%.The study indicates that dimensional reduction and semi-supervised machine learning method is effective in recognizing unsatisfied WeUP in wireless networks.
Highlights
With the development of wireless communications [1]–[14] and internet of things (IoT) [15]–[18], wireless users often require good experience on wireless user perception (WeUP)
Both deep belief autoencoder (DBA) and deep belief network (DBN) consist of a set of Restricted Boltzmann Machines (RBMs) and the network training is the same.The distinguish of DBA is that the number of neurons in each layer is determined by our experimental result. and the number of neurons in core layer is designed as the number of dimensions
The model can be depicted by three parts including data preprocessing, dimensional reduction and semi-supervised clustering
Summary
With the development of wireless communications [1]–[14] and internet of things (IoT) [15]–[18], wireless users often require good experience on wireless user perception (WeUP). Existing DR algorithms perform well in linear relationship between features and face recognition When these DR algorithms are used in the KQI data, the information loss is high and the recognition accuracy is low. Many stateof-the-art semi-supervised methods are developed in wireless communications [34]–[46] Motivated by these previous research, in this paper, we propose an effective WeUP system via dimensional reduction and semi-supervised clustering. This system is distinguished from existing methods. We divide history KQI data into different classes, and find out the most similar class to objects with unsatisfied WeUP by semi-supervised clustering method [48].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
