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

Aprroach for Chance of MT

We often meet difficulties in the calculation of MD. If we meet, we find important hints to solve problems when we consider the Sampling of data.

In Cause and Effect Analysis , the difficulties are the chance to get bif hints to solve problems.

Variance is Zero

Variance of unit space and signal space is zero

We should ignore the variable because the variable is not effect the phenomena.

Variance of unit space is zero. Variance of signal space is not zero

Why the difference of variance?" may be the hint to solve the phenomena.

Multicollinearity

Multicollinearity both unit space and signal space

Only a variable of multicollinearity variables is needed. The others are not needed because multicollinearity means we can not understand the difference of variables.

Multicollinearity of signal space only

The time multicollinearity appears" may be the hint to solve the phenomena.




NEXT Principal Component MT