Abstract
This paper focuses on reducing prediction time for Central Line Associated Blood Stream Infection as one of the main types of Healthcare Associated Infection through a big data analytics model. There is 30,100 Central Line Associated Blood Stream Infection yearly in the US only. It is a severe infection that increases the mortality rate. Big data raises the bar as a result of additional features. It is mainly characterized by a tremendous amount of data that is composed of different forms. It also deals with the rapid data flow rate that is generated from multiple sources, and to top it off the quality of the data is questionable. There has been an increase in the infection rate of HAI during the past few years. Furthermore, the Centers for Disease Control and Prevention updated the definition. Prediction time reduction enables early intervention by clinical staff, which speeds up the recovery time and minimizes harm to the patient. Data mining approach consumes significantly less time, provides higher accuracy, and prevents personal subjective decisions. This paper compares seven data mining algorithms using real patient data of more than 28,000 cases from multiple sources. Naive Bayes shows top accuracy result among other techniques.