A Survey of Data Mining Techniques on Medical Data for Finding Temporally Frequent Diseases

Abstract: Health care domain is flooded with huge amount of data that holds sensitive information pertaining to patients and their medical conditions. Medical data mining can help obtain latent patterns or actionable knowledge. Data mining techniques can discover such latent patterns or hidden relationships among the objects in the medical data sources. This will give know how to ascertain the progression of diseases over a period of time. As medical data sources contain set of observations that are made from time to time with clinical parameters, considering temporal dimension of the data as fundamental parameter can give valuable insights related to temporal nature of diseases. The classical sequence pattern mining is not sufficient to know the temporal nature of diseases that prevail in a region or country. This is because the sequential patterns do not consider the elapsing time between events. Time-annotated sequences can bestow a novel paradigm in data mining. As temporal data mining has potential advantages, this paper focuses on finding data mining techniques that can be used to extract temporally frequent diseases. We analyze the techniques using for temporal data mining on medical data sets.

Keywords: Data mining, medical data mining, data mining techniques, temporally frequent diseases