It is observed, surprisingly, that existing nearest neighbor search methods in wireless data broadcast may not work effectively on mobile clients with very limited memory space. To resolve this problem, a novel method for nearest neighbor search is introduced in the context of a representative of indexes, the grid-partition index, in wireless data broadcast. In the proposed scheme, a mobile client performs the nearest neighbor search by making a sequential access to index packets according to their broadcast order over a wireless channel. The performance evaluation demonstrates that our approach substantially outperforms limited memory versions of existing methods in terms of access time, while retaining a good energy conservation.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Myong-Soo LEE, SangKeun LEE, "Examining Impact of Sequential Access for Nearest Neighbor Search in Wireless Data Broadcast" in IEICE TRANSACTIONS on Communications,
vol. E91-B, no. 9, pp. 2964-2971, September 2008, doi: 10.1093/ietcom/e91-b.9.2964.
Abstract: It is observed, surprisingly, that existing nearest neighbor search methods in wireless data broadcast may not work effectively on mobile clients with very limited memory space. To resolve this problem, a novel method for nearest neighbor search is introduced in the context of a representative of indexes, the grid-partition index, in wireless data broadcast. In the proposed scheme, a mobile client performs the nearest neighbor search by making a sequential access to index packets according to their broadcast order over a wireless channel. The performance evaluation demonstrates that our approach substantially outperforms limited memory versions of existing methods in terms of access time, while retaining a good energy conservation.
URL: https://global.ieice.org/en_transactions/communications/10.1093/ietcom/e91-b.9.2964/_p
Copy
@ARTICLE{e91-b_9_2964,
author={Myong-Soo LEE, SangKeun LEE, },
journal={IEICE TRANSACTIONS on Communications},
title={Examining Impact of Sequential Access for Nearest Neighbor Search in Wireless Data Broadcast},
year={2008},
volume={E91-B},
number={9},
pages={2964-2971},
abstract={It is observed, surprisingly, that existing nearest neighbor search methods in wireless data broadcast may not work effectively on mobile clients with very limited memory space. To resolve this problem, a novel method for nearest neighbor search is introduced in the context of a representative of indexes, the grid-partition index, in wireless data broadcast. In the proposed scheme, a mobile client performs the nearest neighbor search by making a sequential access to index packets according to their broadcast order over a wireless channel. The performance evaluation demonstrates that our approach substantially outperforms limited memory versions of existing methods in terms of access time, while retaining a good energy conservation.},
keywords={},
doi={10.1093/ietcom/e91-b.9.2964},
ISSN={1745-1345},
month={September},}
Copy
TY - JOUR
TI - Examining Impact of Sequential Access for Nearest Neighbor Search in Wireless Data Broadcast
T2 - IEICE TRANSACTIONS on Communications
SP - 2964
EP - 2971
AU - Myong-Soo LEE
AU - SangKeun LEE
PY - 2008
DO - 10.1093/ietcom/e91-b.9.2964
JO - IEICE TRANSACTIONS on Communications
SN - 1745-1345
VL - E91-B
IS - 9
JA - IEICE TRANSACTIONS on Communications
Y1 - September 2008
AB - It is observed, surprisingly, that existing nearest neighbor search methods in wireless data broadcast may not work effectively on mobile clients with very limited memory space. To resolve this problem, a novel method for nearest neighbor search is introduced in the context of a representative of indexes, the grid-partition index, in wireless data broadcast. In the proposed scheme, a mobile client performs the nearest neighbor search by making a sequential access to index packets according to their broadcast order over a wireless channel. The performance evaluation demonstrates that our approach substantially outperforms limited memory versions of existing methods in terms of access time, while retaining a good energy conservation.
ER -