I am a beginner in statistics, therefore I hope I can state my problem in a correct manner. I have a some instances or samples and I can collect below statistical parameters for **classification and regression** problem:

- Sample Size
- Minimum value
- Maximum value
- Standard deviation
- Variance
- Mean

And, I want to use **z-scores** to compare or classify samples, my question is: does using z-score **make sense** or what can I use instead of z-score to obtain meaningful classification parameter?

**Contents**hide

#### Best Answer

In order to perform linear regression you'd need not only the means and variances of the variables but also all their covariances (or equivalently their correlations).

If you can collect the means, variances and covariances/correlations separately in each of the classes you wish to classify, you can do linear discriminant analysis, which is a classification method, albeit a somewhat old-fashioned one with some rather restrictive normality assumptions.

### Similar Posts:

- Solved – Why is Pearson’s ρ only an exhaustive measure of association if the joint distribution is multivariate normal
- Solved – Classification with confidence scores: is regression ok
- Solved – Classification with confidence scores: is regression ok
- Solved – Regression: Covariance table in spss vs. mplus
- Solved – Is it possible to use Recurrent NN (LSTM) for classification