Wyniki wyszukiwania - BazTech

1

Location-based Power Control Mechanism for D2D Communication Underlaying a Cellular System

Rodziewicz Marcin

Journal of Telecommunications and Information Technology

|

2023

|

nr 3

49--53

EN

The paper presents a location-based approach to controlling the power of device-to-device (D2D) underlay of a frequency reuse-1 cellular system. The system allows for direct communication to share uplink resources with cellular users. As a result, both D2D and cellular users are experiencing additional interferences in the system. By controlling the output power of the devices, these interferences can be mitigated and the performance of the network can be improved in terms of better spectral and energy efficiency. The proposed location-based target signal-to-interference ratio power control scheme for D2D communications utilizes information about users' locations to estimate the interference level experienced by the receiver of the direct link. Based on this estimation, an appropriate transmit power can be determined. The performance of the proposed power control solution is investigated via system level simulations.

2

Using contextual conditional preferences for recommendation tasks: a case study in the movie domain

Karpus A., Di Noia T., Tomeo P., Goczyła K.

Studia Informatica

|

2016

|

Vol. 37, nr 1

7--18

EN

Recommendation engines aim to propose users items they are interested in by looking at the user interaction with a system. However, individual interests may be drastically influenced by the context in which decisions are taken. We present an attempt to model user interests via a set of contextual conditional preferences. We show that usage of proposed preferences gives reasonable values of the accuracy and the precision even when the dataset is quite small.

PL

Systemy rekomendacyjne sugerują użytkownikom produkty, którymi mogą być zainteresowani, na podstawie wcześniejszej interakcji z systemem. Jednak duży wpływ na decyzję użytkownika ma kontekst, w którym jest ona podejmowana. W artykule zaproponowano model zainteresowań użytkownika jako zbiór kontekstowych preferencji warunkowych i pokazano, że z ich wykorzystaniem można uzyskać dużą dokładność i precyzję rekomendacji, nawet dla małych zbiorów danych.

3

On using data mining techniques for context-aware student grouping in e-learning systems

Zakrzewska D.

Information Systems in Management

|

2014

|

Vol. 3, No. 1

77--88

EN

Performance of an e-learning system depends on an extent to which it is adjusted to student needs. Priorities of the last ones may differ in accordance with the context of use of an e-learning environment. For personalized e-learning system based on student groups, different distribution of the groups should be taken into account. In the paper, using of data mining techniques for building student groups depending on the context of the system use is considered. As the main technique unsupervised classification is examined,. Context parameters depending on courses and student models are tested. Experiment results for real student data are discussed.