<![CDATA[This study aims to predict the processing time of the approval of Clinical Practice Guidelines (CPG), which exhibits considerable variation in duration and is difficult to predict accurately. In addition, the utilization of historical hospital administrative data to build effective predictive models for estimating the duration of the CPG approval process has not yet been optimized. Therefore, this research seeks to develop a predictive model to estimate the processing time of the CPG approval process.The proposed approach employs deep learning techniques by leveraging historical administrative data as the basis for modeling. The methods applied include K-Means Clustering, Decision Tree, and Long Short-Term Memory (LSTM). K-Means Clustering is used to group CPG data based on similar administrative characteristics, enabling the identification of approval time patterns. Subsequently, the Decision Tree method is utilized to analyze the relationships among variables and to generate classification rules that explain the factors influencing the duration of the CPG approval process. Meanwhile, LSTM serves as the primary model for predicting the processing time of CPG approval.This study uses 487 CPG records collected over the period from 2020 to 2024. The evaluation results indicate that the K-Means Clustering method achieves an accuracy rate of 87,36%. This level of accuracy reflects strong clustering performance and a high degree of conformity with actual conditions, indicating that the results are suitable to be used as a foundation for further analysis in the classification and prediction stages of the CPG approval process.]]>
