Anita Sindar Ros Maryana Sinaga, Ricky Eka Putra
There is a wealth of data that can be used to improve public health services by leveraging machine intelligence technology. The problem that arises is the distribution of unstructured data, not detailed and incomplete. In this study, data is processed with the machine learning algorithm classification stages that are tailored to the case and the objectives to be achieved. This Big Data Analytics model uses Descriptive analytics in combination with Predictive Analytics to work by looking at future situations in relation to past data. Prescriptive Analytics tries to analyze decision making to deal with this condition. Predictive analytics aims to analyze what patterns emerge in the health sector to improve health services. The Naive Bayes classifier is used to classify whether a patient will return for treatment or not, based on the village area the probability of returning to treatment is 0.40 and not returning to treatment is 0.60. Of the 13 health indicators, the data are grouped into 2 classes, namely fulfilled and unfulfilled by using a type of algorithm that includes the SVM Classifier. The model is same to have perfect prediction accuracy if the AUC value is 0.90. The best test model is Coarse Gaussian SVM result 90.1%. Grouping through decision trees aims to inform the adequacy of tree forecast accuracy = 97.9%, Linear Regression = 96.6%. © 2022 IEEE.
Information Technology, Stmik Pelita Nusantara, Medan, Indonesia; Informatics Engineering, Surabaya State University, Surabaya, Indonesia
