Missing Data Problem in Predictive Analytics

A revolution in computational methods and statistics to process and analyse data into insight and knowledge is along with the growth of data. The paradigm of data analytic is changed from explicit to implicit raises the way to extract knowledge from data through a prospective approach to determine the value of new observations based on the structure of the relationship between input and output (predictive analytics). In the cycle of predictive analytics, data preparation is a very important stage. The main challenge faced is that raw data cannot be directly used for analysis and related to the quality of the data. Completeness is arising related to data quality. Missing data is one that often causes data to become incomplete. As a result, predictive analytics generated from these data becomes inaccurate. In this paper, the issues related to the missing data in predictive analytics will be discussed through a literature study from related research. Also, the challenges and direction that might occur in the predictive analytics domain with problems related to missing data will be presented.