10楼: Originally posted by feixiaolin at 2014-12-04 20:56:17
你是三坐标机，摄影测董，还是卡规？
你的测法中有一个端点很难判读。
...

12楼: Originally posted by zaq123321 at 2014-12-05 10:35:23
I think the general idea should work. You said you have figured out the reason why it conflicts with what you expected. You may first post here for others' reference.  I have several confusions. Firs ...

12楼: Originally posted by zaq123321 at 2014-12-05 10:35:23
I think the general idea should work. You said you have figured out the reason why it conflicts with what you expected. You may first post here for others' reference.  I have several confusions. Firs ...

13楼: Originally posted by hqusurui at 2014-12-05 10:56:30
十分感谢，我必须要解释一下，这一个样片中存在的8个数据，他们是有某种关系的，由于加工水平的原因，两个比较大，两个比较小，所以导致了我们同时统计800个片的话，把极端的数据聚集在5%的片上，实际上这些极端的 ...

15楼: Originally posted by zaq123321 at 2014-12-05 11:20:17
If I understand correctly, in these 8 data. Two bigger, two smaller, 4 in middle. Just use one normal distribution is not good. Three normal distributions should be better. One group has 200 data, a ...

16楼: Originally posted by hqusurui at 2014-12-05 11:34:22
这个我们想过了，我会试试，没有更加高端的方法吗？...

17楼: Originally posted by zaq123321 at 2014-12-05 11:45:35
One normal distribution is not enough since the data come from different sampling. Their ensambles are not the same. How can you just use one normal distribution to represent them?...