Page 369 - Special Topic Session (STS) - Volume 4

STS1080 Mahmoud Rafea et al.
employed to denote information systems in which some symbolic
representation of human knowledge of a domain is applied, usually in a way
resembling human reasoning, to solve actual problems in the domain. As this
knowledge is often derived from experts in a particular field, and early
knowledge-based systems were actually developed in close collaboration with
experts. However, the new direction is extracting the knowledge and storing
it to facilitate the decision. Rules are the most appropriate knowledge in
medical problems. Thus, association rule mining algorithms play a vital role in
solving these problems especially in disease diagnosis.
Some types of diseases are difficult to detect at early stage due to the lack
of symptoms. Early detection of serious diseases is essential in reducing life
losses. Earlier treatment, however, requires the ability to detect these diseases
in early stages. Early diagnosis requires accurate and reliable diagnosis
procedure. Automatic diagnosis is considered as a real world medical problem.
Therefore, finding an accurate and effective diagnosis method is very
important [4]. Thus, the new direction in disease diagnosis is based on
proteomics. This is by using machine learning on proteomics data to extract
biomarkers, and to identify the serious diseases.
The Association Rule Mining on Medical Applications is described in
section 2. The Biomarkers Discovery is described in section 3. Applying
machine learning algorithms on EDAS are described in section 4. Conclusion
and future work are described in section 5.

2. Association Rule Mining on Medical Applications
Multiple papers reviewed the solution of medical problems as Association
Rules. In [5], the researchers proposed a data mining technique based on
Apriori Algorithm for generating the frequency of diseases that affect patients
in the various geographical region and at various time periods. The analyses
concluded that patients are affected by 4 different diseases in a particular
geographical area during a particular year.
In [6], the researchers presented an association rule mining for medical
data to anticipate heart diseases using Apriori algorithm. They used medical
data which the diseased and healthy patient’s details are categorized for the
prediction of heart diseases.
In [7], it was focused on the implementation of the Apriori Algorithm to
discover interesting patterns and association rules in chronic diseases.
Percentage of possibility for chronic disease was calculated from each
symptom of all considered chronic diseases. The higher number of symptoms
lead to higher accuracy of calculating the disease possibility.
However, in [8], the researchers selected the topic of NED (No Evidence of
Disease) and ED (Evidence of Disease) for the Breast Cancer problem. They
experimented two association rule mining algorithms; Apriori and FP-Growth.
They attempted to detect the relationships between different factors. Their

358 | I S I W S C 2 0 1 9