Dot leicester organization

Many companies use data mining processes to discover the behavioral patterns of their customers. However, the mining process is usually only done on a sample of the overall data. This comes with a little problem build in, i.e. what if the sample that was used was not representative of the whole database or data warehouse? Can you really rely on data and information like that?

Some people will say that data mining on mere samples of data leaves you too open to missing the “Big Picture.” They will say that it is too easy to extract a sample that doesn’t include a hint of the patterns that may exist in the overall data structure. However, the solution is simple. The answer is that you can rely on samples, but not on a single sample. Always use other samples to verify your original results and make sure you have a representative sample of information and data.