att_abstract={{Localized user density estimation is groundwork in many fields as urban planning, traffic engineering, disease control, location based marketing and telecomm capacity planning.  Modern mobility technologies provide the capability for measuring the user density dynamically and precisely. But this only limits to the areas that have good signal strength. It is a challenge to estimate the user density accurately for the areas with poor signal strength, Especially the small areas.  Benefit from the all perspective big data collected by telecomm, this paper proposes a method to estimate user density for small areas with  poor single strength.  The proposed method segment a geographic area into small grids with good signal coverage and bad signal coverage. The user density for bad coverage grids is inferred from the good coverage grids with similar geographic, demographic and  firmographic information base on the K-Nearest-Neighbor concept. Instead of predefining the K, different percentile measurements are provided to increase the robustness in capacity planning. }},
	att_authors={rd1424, gm1461},
	att_copyright_notice={{This version of the work is reprinted here with permission of IEEE for your personal use. Not for redistribution. The definitive version was published in 2014. {{, 2014-05-09}}
	author={Rong Duan and Guang-qin Ma},
	institution={{Proceeding of the 23rd Wireless and Optical Communication Conference(WOCC 2014) 
May 9-10 2014, NJI}},
	title={{Adjusted KNN Model in Small Area User Density Estimation with Poor Signal Strength }},