here is the right way to analyze data


所有跟贴·加跟贴·新语丝读书论坛

送交者: 008 于 2012-02-07, 16:43:03:

回答: 程旭辉:截至目前方舟子vs韩寒 调查结果对比 由 flyfreely 于 2012-02-07, 12:04:21:

I have published this theory before on international journal xys, but nobody paid attention. Here I repeat it again. Hope it can help you understand the statistic of F-fan and H-fan better.

result = sum_i( sign(person i) * (how much you care for his/her opinion) * his opinion )

basically, you summarize people's opinion with a weight factor that you care or not care for his/her opinion. However, the most important part is the sign. Some people has a positive weight, some has a negative weight. You determine the sign by your own judgement.

For example, FZZ has a weight of 10, Hunhunseng has a weight of 8, abada has a weight of -5.

For the statement of "HH is a fake author", Fzz thinks it is positive, but abada thinks it is negative. Hunhunseng is absent.

result = 10 x 1 + 8 x 0 + (-5) x (-1) =15.

It is a very strong suggestion that the statement is true.

Now you see a simple comparison of the numbers of F-fan and H-fan is meaningless. You have to watch out the weight factors. Most importantly, pay attention to the signs!




所有跟贴:


加跟贴

笔名: 密码: 注册笔名请按这里

标题:

内容: (BBCode使用说明