Skip to main content

Table 11 Accuracy evaluation results - Varying the value of k test

From: Efficient spatial data partitioning for distributed \(k\)NN joins

  K Exact match Mis-match Missing Exact match Mis-match Missing
With non-spatial data* Without non-spatial data*
SpPart_kNN  VS   LocationSpark
BUS Records (119, 319) 3 \(91.19\%\) \(8.81\%\) \(0.01\%\) \(90.86\%\) \(9.14\%\) \(0.01\%\)
10 \(81.72\%\) \(18.28\%\) \(0.01\%\) \(80.60\%\) \(19.40\%\) \(0.01\%\)
50 \(30.00\%\) \(70.00\%\) \(0.01\%\) \(25.46\%\) \(74.54\%\) \(0.01\%\)
100 \(10.17\%\) \(89.83\%\) \(0.01\%\) \(5.64\%\) \(94.36\%\) \(0.01\%\)
500 \(0.01\%\) \(99.99\%\) \(0.01\%\) \(0\%\) \(100.00\%\) \(0.01\%\)
1000 \(0\%\) \(100.00\%\) \(0.01\%\) \(0\%\) \(100.00\%\) \(0.01\%\)
TAXI Records (119, 319) 3 \(95.31\%\) \(4.69\%\) \(0\%\) \(95.29\%\) \(4.71\%\) \(0\%\)
10 \(81.95\%\) \(18.05\%\) \(0\%\) \(82.02\%\) \(17.98\%\) \(0\%\)
50 \(49.19\%\) \(50.81\%\) \(0\%\) \(49.06\%\) \(50.94\%\) \(0\%\)
100 \(38.64\%\) \(61.36\%\) \(0\%\) \(38.44\%\) \(61.56\%\) \(0\%\)
500 \(26.42\%\) \(73.58\%\) \(0\%\) \(25.98\%\) \(74.02\%\) \(0\%\)
1000 \(15.61\%\) \(84.39\%\) \(0\%\) \(14.31\%\) \(85.69\%\) \(0\%\)
TLC Records(119, 319) 3 \(96.66\%\) \(3.34\%\) \(0\%\) \(96.67\%\) \(3.33\%\) \(0\%\)
10 \(85.67\%\) \(14.33\%\) \(0\%\) \(85.61\%\) \(14.39\%\) \(0\%\)
50 \(33.30\%\) \(66.70\%\) \(0\%\) \(33.00\%\) \(67.00\%\) \(0\%\)
100 \(17.09\%\) \(82.91\%\) \(0\%\) \(16.85\%\) \(83.15\%\) \(0\%\)
500 \(0.25\%\) \(99.75\%\) \(0\%\) \(0.18\%\) \(99.82\%\) \(0\%\)
1000 \(0.01\%\) \(99.99\%\) \(0\%\) \(0.00\%\) \(100.00\%\) \(0\%\)
  1. *Simba, STARK, GeoSpark results omitted for lack of support or exceeding 180 min of runtime