ADVANCES BUSINESS ANALISIS
Introduction :
We can answers to what happene,d why, what is happening. But difficlut to answer what will happen ? and we can often discover uncexpected connection in the business !
Data mining is defi ned as “the nontrivial extraction of implicit, previously unknown, and potentially useful information (patterns) from data.” This is called knowledge discovery. The most important thing is to identify the patterns, whcich allow us to deine the structure ;
We can say tjat data mining gives us knowledge.
The most common application of data mining are : classification, prediction, cluster analisis (objetcs that have similars featurs) , mining association rules

1) Classification : trees
Ex : the decision to grant credit
How you construct a classifier : learning on the basis of learning eexamples (examples of correctly categorized objects) it gives us learning system (algorytms) and then classifier.
Limitations : To construct a classifi er on the basis of a set of examples, you need to solve many problems that are common for the majority of data-mining algorithms. However, if you are aware of these limitations, you should have reasonable expectations regarding their possible applications and the quality of the knowledge generated by them.
The main problems are connected with induction, history, updating, and overfi tting : * Induction problem : learning from examples is inductive reasoning : so we make generalizacion, from limited observation. And the limit is that we are not able to prove that an inductive conclusion is true. Examples can be wrong, and even when they are true, inductive conslusion are incertain ! * History and updating : generzting a classifer implies Learning on the basis of historical observations. But reality is subject to changes in time ! this produce uncertainty of classification. Classifer must be