Hi,
I have a set of data (total number of record = 144,122), and I would like to
use gamma-glm with log link to set up a model.
IC is number of records
IL is paid amount
The table below shows that I have
30.578% of the data in the level of "1 - 1000" paid amount
20.320% of the data in the level of "1001 - 2000" paid amount
and so on
My question is could i use the whole data set to model ?
or may be i have to use the data up to 10,000 paid amount ?
Level IL Avg IL Sum IC 1-1000 539.60 44,069 30.578% 1001
- 2000 1,444.81 29,285 20.320% 2001 - 3000 2,457.72 15,343
10.646% 3001 - 4000 3,473.40 8,497 5.896% 4001 - 5000 4,496.47
5,838 4.051% 5001 - 6000 5,476.28 3,831 2.658% 6001 -
7000
6,482.82 2,889 2.005% 7001 - 8000 7,500.97 2,323 1.612% 8001
- 9000 8,492.07 1,772 1.230% 9001 - 10000 9,542.60 1,736
1.205% 10001 - 11000 10,490.19 1,291 0.896% 11001 - 12000 11,516.00
1,104 0.766% 12001 - 13000 12,508.65 915 0.635% 13001 -
14000 13,501.24 869 0.603% 14001 - 15000 14,562.99 876
0.608% 15001 - 16000 15,502.50 650 0.451% 16001 - 17000 16,498.90
585 0.406% 17001 - 18000 17,511.23 573 0.398% 18001 -
19000 18,529.04 512 0.355% 19001 - 20000 19,605.73 615
0.427% 21001 - 22000 21,518.71 448 0.311% 22001 - 23000 22,489.74
389 0.270% 23001 - 24000 23,493.52 340 0.236% 24001 -
25000 24,603.33 413 0.287% 25001 - 26000 25,516.57 324
0.225% 26001 - 27000 26,514.70 297 0.206% 27001 - 28000 27,509.62
272 0.189% 28001 - 29000 28,486.77 238 0.165% 29001 -
30000 29,591.02 312 0.216% 30001 - 31000 30,460.08 238
0.165% 31001 - 32000 31,527.67 240 0.167% 32001 - 33000 32,526.25
213 0.148% 33001 - 34000 33,496.41 208 0.144% 34001 -
35000 34,556.44 235 0.163% 35001 - 36000 35,476.62 190
0.132% 36001 - 37000 36,512.92 155 0.108% 37001 - 38000 37,524.77
191 0.133% 38001 - 39000 38,469.21 152 0.105% 39001 -
40000 39,554.93 176 0.122% 40001 - 41000 40,501.28 171
0.119% 41001 - 42000 41,521.06 182 0.126% 42001 - 43000 42,525.54
156 0.108% 43001 - 44000 43,541.32 118 0.082% 44001 -
45000 44,549.38 131 0.091% 45001 - 46000 45,513.78 125
0.087% 46001 - 47000 46,532.88 128 0.089% 47001 - 48000 47,528.92
121 0.084% 48001 - 49000 48,472.52 115 0.080% 49001 -
50000 49,684.14 191 0.133% 50001 - 51000 50,556.99 104
0.072% 51001 - 52000 51,527.56 119 0.083% 52001 - 53000 52,519.82
120 0.083% 53001 - 54000 53,504.29 105 0.073% 54001 -
55000 54,527.36 126 0.087% 55001 - 56000 55,566.20 94
0.065% 56001 - 57000 56,533.86 98 0.068% 57001 - 58000
57,494.78
112 0.078% 58001 - 59000 58,555.27 100 0.069% 59001 -
60000 59,592.80 134 0.093% 60001 - 61000 60,540.90 94
0.065% 61001 - 62000 61,488.11 93 0.065% 62001 - 63000
62,543.60
95 0.066% 63001 - 64000 63,509.94 95 0.066% 64001 -
65000 64,578.59 97 0.067% 65001 - 66000 65,486.95 82
0.057% 66001 - 67000 66,518.80 72 0.050% 67001 - 68000
67,507.95
85 0.059% 68001 - 69000 68,516.98 86 0.060% 69001 -
70000 69,541.52 89 0.062% 70001 - 71000 70,519.49 77
0.053% 71001 - 72000 71,488.88 74 0.051% 72001 - 73000
72,483.43
72 0.050% 73001 - 74000 73,489.59 82 0.057% 74001 -
75000 74,601.14 92 0.064% 75001 - 76000 75,467.29 78
0.054% 76001 - 77000 76,550.53 74 0.051% 77001 - 78000
77,525.18
76 0.053% 78001 - 79000 78,481.88 74 0.051% 79001 -
80000 79,555.44 78 0.054% 80001 - 81000 80,443.24 72
0.050% 81001 - 82000 81,504.18 67 0.046% 82001 - 83000
82,510.68
78 0.054% 83001 - 84000 83,483.89 78 0.054% 84001 -
85000 84,566.99 84 0.058% 85001 - 86000 85,504.36 86
0.060% 86001 - 87000 86,584.49 67 0.046% 87001 - 88000
87,540.17
60 0.042% 88001 - 89000 88,483.13 73 0.051% 89001 -
90000 89,532.23 70 0.049% 90001 - 91000 90,521.94 73
0.051% 91001 - 92000 91,597.56 62 0.043% 92001 - 93000
92,499.65
75 0.052% 93001 - 94000 93,515.79 64 0.044% 94001 -
95000 94,524.84 76 0.053% 95001 - 96000 95,469.49 81
0.056% 96001 - 97000 96,492.13 55 0.038% 97001 - 98000
97,454.56
62 0.043% 98001 - 99000 98,493.23 57 0.040% 99001 -
100000 99,593.13 90 0.062%
IL > 100,000 is 6.85%
Thank you
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
