Skip to content
SDGs Center UDINUS

Evidence record · UDINUS · SDGs

An Optimally Configured Hybrid Model for Healthcare Time Series Prediction

Rektorat UDINUS 2011 Published 2026-10-30

Evidence record · UDINUS · SDGs

Unit
Rektorat UDINUS
Period
2011
Goals
Record
7f45a248-040d-48ca-978f-d662e461438a
Published
2026-10-30

Universitas Dian Nuswantoro conducted an activity involving researchers Purwanto, Eswaran, Chikkannan, and Logeswaran, Rajasvaran to address the challenge of improving accuracy in healthcare time series prediction. The activity focused on healthcare data, including mortality and morbidity data, which are described as health indicators of a society and as relevant to healthcare management.

The activity examined limitations in the use of linear prediction models, including autoregressive integrated moving average and linear regression. These models were described as having limited ability to handle nonlinear relationships among data. Neural network models were considered more suitable for handling such relationships. However, healthcare time series data contain complex linear and nonlinear patterns, making it difficult to obtain high prediction accuracy by using only linear models or only neural network models.

UDINUS researchers proposed a hybrid method that combines the best linear model with an optimally configured neural network. Unlike other hybrid models that use predetermined configurations for their linear and neural network components, this method selects the best linear model and the optimum neural network configuration according to the type of input data. The proposed method was tested using two types of healthcare data: infant mortality rate data and morbidity of malaria data.

The experiment results showed that the proposed hybrid model yielded more accurate prediction results than other known models. In relation to the stated goal, the activity contributes to health research by addressing prediction accuracy for healthcare time series data through a method that considers both linear and nonlinear patterns. The activity provides an institutional example of UDINUS research addressing methodological challenges in healthcare data prediction.

Reference https://www.scopus.com/pages/publications/80055040525?origin=resultslist