Fine-grained Association Rules toward knowledge discovery

One of the basic problems in data processing is prediction. To predict, we need to extract valuable information out of the unstructured data. The value of information is how easy we can access and retrieve information. Data mining techniques as well as knowledge discovery methodologies aim at discovering hidden patterns in unstructured mass of data. In this paper, I propose the sequence of actions taken toward extraction of valuable information out of a salary dataset by selecting some rules out of the association rules through calculation of the support and confidence associated with them. Finally in order to normalize the published association rules, a merging of the rules is implemented and new information will be extracted accordingly. The result of this process will be association rules, which can be considered as meaningful and formal rules, applicable on the chosen dataset.