Verfahren, Computerprogramme und Gerät zum Trainieren eines Maschinenlernmodells

Details

Veröffentlichungs-Nr.

EP4746318

Anmeldetag

15. November 2024

Veröffentlichung

20. Mai 2026

Rechtsraum

EP

Abstract

Examples relate to methods, computer programs, and apparatuses for training a machine learning model to predict a radio quality for user equipment, UE, at a location covered by a mobile communication system and for predicting a radio quality for user equipment at a location covered by a mobile communication system. The method (10) for training a machine learning model to predict a radio quality for user equipment, UE, at a location covered by a mobile communication system comprises generating (11) training data with basic feature training sets, which are based on measurements conducted at positions in a serving radio cell, additional feature sets comprising information about radio parameters of the serving radio cell and/or one or more neighboring radio cells, geometrical features of measurement positions, wherein the geometrical features comprise distances being determining for a radio propagation channel between the serving radio cell and a measurement position, and radio qualities obtained at measurement positions. The method (20) further comprises training (12) the machine learning model with the training data.

Anmelder

Firma

Deutsche Telekom AG

Land

🇩🇪 Deutschland