Abstract
Malignant Tumors are developed over several years due to unknown biological factors. These biological factors induce changes in the body and consequently, they lead to Malignant Tumors. Some habits and behaviors initiate these biological factors. In effect, the immune system cannot recognize a Malignant Tumor as foreign tissue. In order to discover a fascinating pattern of these habits, behaviors, and diseases and to make effective decisions, different machine learning techniques should be used. This research attempts to find the association between normal proteins (environmental factors) and diseases that are difficult to diagnose and propose justifications for those diseases. This paper proposes a technique for medical data mining using association rules. The proposed technique overcomes some of the limitations in current association algorithms such as the
Apriori
algorithm and the Equivalence CLAss Transformation (ECLAT) algorithm. A modification to the
Apriori
algorithm has been proposed to mine Erythrocytes Dynamic Antigens Store (EDAS) data in a more efficient and tractable way. The experiments inferred that there is a relation between normal proteins as environment proteins, food proteins, commensal proteins, tissue proteins, and disease proteins. Also, the experiments show that habits and behaviors are associated with certain diseases. The presented tool can be used in clinical laboratories to discover the biological causes of malignant diseases.