Technical practice abounds with numerous diverse data records. Sometimes the data is complete, sometimes it is censored or truncated. It is not always easy and straightforward to record the data. And even after, the data processing is by no means simple, especially when the data forms a significant set of a huge size and large informational diversity. Typically, the data containing more observed variables, either dependent or independent, is called multidimensional. Also, if the multidimensional data contains numerous records, it is not easy to determine which dependent or independent variables are important for further study. Our aim and ambition is to introduce a couple of methods which are very suitable and sometimes absolutely necessary for exploratory data analysis. The methods help us to determine i) the level of significance of the data for single recorded variables, ii) the level of mutual dependence among the data, and iii) the choice of the best representatives for further data study. The recommended methods used for the exploratory data analysis are presented with practical examples.
展开▼
