Big Data Analysis and Data Mining

Big Data leads to the acceptance that based on the large amount of data, the true error rate can be estimated without trouble. However, if this true has be proofed. Large data might not be well distributed in the solution space and subareas of the solution space might be overrepresented than other subareas. Therefore, it is important to know the true situation in the data sample.

We review in the talk what can be achieved with sampling, how the true error rate is estimated and what the measures that are calculated during the estimation of the error rate tell us about the true situation.