The analytical results induced fifteen representative frequent patterns, these frequent patterns are consists with physicians domain knowledge, and are found very helpful to the existing clinical treatments. Furthermore, compared to the traditional Apriori algorithm, extra 3 to 4 frequent patterns are found via our proposed analysis architecture on average. In additional, compared the performance of stand-alone and map reduce architecture of the proposed architecture, the implementation of our proposed method with three reduces was 12.3% faster in terms of the implementation-time efficiency. Our method may prove extremely helpful in improving doctor-patient relationships and overall health care quality.
Journal of Medical Imaging and Health Informatics, Vol.7, No.6, pp.1399-1408