Top Page | Upper Page | Contents | About This Site | JAPANESE

Errors

If we know errors, we know the reason of the case that our data analysis goes bad.

And if we know, our analysis becomes strong.

Kinds of Errors

Errors of measuring are factors of unevenness of quality . There are accidental errors and systematic errors. In actual data, the border between them is not clear.

If the management of measuring becomes better, the error that we think "accidental" changes "systematic" and we may control the errors.

Accidental Errors

We can find accidental errors when we measure same object in a same way repeatedly. The unevenness is shown as normal distribution .

If we measure once, we cannot estimate the size of the accidental error. And we cannot estimate the rightness of the data.

Systematic Errors

Systematic errors are made by the tendency of the people and the machine of measuring.

If there are not any systematic errors, average is the nearest number for the right number. But systematic errors prevent to estimate the right number.

Know-hows, experiences and intuition are important to estimate systematic errors.

Traditional statistics is a study for the condition that there are not systematic errors.

Noise

Noise is the obstacle of signal. In electrical engineering, "SN ratio" is an index to measure the size of noise. SN stands for signal and noise. SN ratio is calculated that the voltage of noise is divided by the voltage of noise.

In Quality Engineering , we see "SN ratio". But SN Ratio of quality engineering is different from that of electrical engineering. Quality engineering defines the noise as errors.




NEXT Measurement Errors

Tweet