Open AccessRandom forests‐enabled context detections for long‐term evolution network forrailway
Author(s) -
Zhang Lei,
Ni Qin,
Zhang Guanglin,
Zhai Menglin,
Moreno Juan,
Briso Cesar
Publication year - 2019
Publication title -
iet microwaves, antennas and propagation
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.555
H-Index - 69
eISSN - 1751-8733
pISSN - 1751-8725
DOI - 10.1049/iet-map.2018.6025
Subject(s) - random forest , support vector machine , context (archaeology) , term (time) , computer science , linear discriminant analysis , decision tree , artificial intelligence , machine learning , discriminant , construct (python library) , tree (set theory) , ensemble learning , pattern recognition (psychology) , data mining , mathematics , geography , computer network , mathematical analysis , physics , archaeology , quantum mechanics
An ensemble learning‐based approach for context detection for high‐speed railway (HSR) isproposed, evaluated, and compared against various machine learning algorithms.The context information is collected and formatted from the realistic physicaldata, which are measured by the field test in a commercial 4G cellular networkon‐board high‐speed train in an open area, HSR station, and typical urban area.The radio propagation knowledge implies that the channel model is implemented infeature extractions. The independent out‐of‐bag errors show ensemble learningapproaches; especially, random forests‐based algorithm achieves very accuratecontext detection (up to 93.5%), which is much higher than single tree (66%).Other ensembles with sub‐spaces of discriminant (67%) and K‐nearest neighbour(69%), as well as linear discriminant analysis (55.7%), and support vectormachine (SVM) (27.7%). Furthermore, the features are selected based on thefeature importance evaluation in the first round training. The selectedpredominant features and the reasonable number of individual trees construct therefined random forests, which obtain a high accuracy (92.6%) and 70% timereducing. This experimental study benefits from physical radio knowledge tosupport advanced long‐term evolution (LTE) network for railway and future smartrail applications.