[原创]计量版人气与回复的回归分析：）

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

取计量版2006年1月19日20：20数据

回复	人气
85	900
7454	92435
112	8945
78	1583
10	27511
173	9931
865	23010
416	15947
17	339
300	10674
182	5767
8	26278
16	18510
79	2517
2	11
2	14
0	8
48	846
1	8
0	7
2	15
0	4
0	3
0	7
15	7
214	218
6	6677
0	119
0	7

[此贴子已经被作者于2006-1-20 20:56:43编辑过]

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：回归分析 人气 计量 原创 回归分析

样本空间为40

第二步：利用散点图确认可使用线性模型

第三步：取人气为解释变量

回复为因变量

进行回归分析

[此贴子已经被作者于2006-1-20 20:37:40编辑过]

SUMMARY OUTPUT
回归统计
Multiple R	0.904042
R Square	0.817292
Adjusted R Square	0.810525
标准误差	600.0615
观测值	29
方差分析
	df	SS	MS	F	Significance F
回归分析	1	43488488	43488488	120.7766	1.81E-11
残差	27	9721993	360073.8
总计	28	53210481
	Coefficients	标准误差	t Stat	P-value	Lower 95%	Upper 95%	下限 95.0%	上限 95.0%
Intercept	-247.038	123.8772	-1.99422	0.056317	-501.213	7.137038	-501.213	7.137038
人气	0.068368	0.006221	10.98984	1.81E-11	0.055603	0.081132	0.055603	0.081132
RESIDUAL OUTPUT					PROBABILITY OUTPUT
观测值	预测 回复	残差	标准残差		百分比排位	回复
1	-185.507	270.5067	0.459071		1.724138	0
2	6072.554	1381.446	2.344418		5.172414	0
3	364.5135	-252.513	-0.42853		8.62069	0
4	-138.811	216.8114	0.367946		12.06897	0
5	1633.833	-1623.83	-2.75577		15.51724	0
6	431.9243	-258.924	-0.43941		18.96552	0
7	1326.109	-461.109	-0.78254		22.41379	0
8	843.2259	-427.226	-0.72503		25.86207	1
9	-223.861	240.8612	0.40876		29.31034	2
10	482.7217	-182.722	-0.31009		32.75862	2
11	147.2401	34.75989	0.05899		36.2069	2
12	1549.535	-1541.54	-2.6161		39.65517	6
13	1018.453	-1002.45	-1.70124		43.10345	8
14	-74.9558	153.9558	0.261275		46.55172	10
15	-246.286	248.2859	0.42136		50	15
16	-246.081	248.0808	0.421012		53.44828	16
17	-246.491	246.491	0.418314		56.89655	17
18	-189.199	237.1986	0.402544		60.34483	48
19	-246.491	247.491	0.420011		63.7931	78
20	-246.559	246.5593	0.41843		67.24138	79
21	-246.012	248.0124	0.420896		70.68966	85
22	-246.764	246.7644	0.418778		74.13793	112
23	-246.833	246.8328	0.418894		77.58621	173
24	-246.559	246.5593	0.41843		81.03448	182
25	-246.559	261.5593	0.443886		84.48276	214
26	-232.134	446.1337	0.757123		87.93103	300
27	209.455	-203.455	-0.34528		91.37931	416
28	-238.902	238.9021	0.405435		94.82759	865
29	-246.559	246.5593	0.41843		98.27586	7454

得出回归方程：y=-247.038+0.068368x

Multiple R 0.904042
R Square 0.817292

拟和状况和相关状况良好

[此贴子已经被作者于2006-1-20 20:57:39编辑过]

so I think you may add other explanatory variable

your constant may contains risk premium

why,please tell me your concrete reasons

pleased to communicate with u and anyone who likes to study statistics.

maybe someone can chang it better

it is interesting

yes, the constant contains risk premium,maybe the sample is too centralized.

because people have different preference for replying it

some people may like to input water

some people may be so lazy that never reply

hence add risk premium is a good choice

平均每1000个读帖者中只有约436.642个回帖

结论：潜水者众，回帖者少

建议：大家每个人都更加preactive一些，多多参与，促进论坛的发展。

what my suggestion is you may put dummy variable

d=1, valuable reply

d=0 water