Advanced forecasting of maternal mortality rates: A comparative analysis of ARIMA, LSTM, and hybrid ARIMA-LSTM models
Nurshazwani Muhamad Mahfuz, Hazlina Haron, Noor Hafizah Ismail
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
High rates of maternal mortality are still among the global threats of our age, especially in developing countries such as Malaysia. Specifically, this paper aims to predict and compare the future maternal mortality of traditional ARIMA, LSTM, and hybrid LSTM-ARIMA models. The hybrid of the LSTM-ARIMA model further improves the predictive capabilities that can be indicated by traditional ARIMA and the high pattern recognition ability of LSTM. The original data set of historical maternal mortality was pre-processed by removing any outliers, then standardized, and then split into train and test data sets. Before training the predictive model, the linear trends of the data were removed through the use of the ARIMA model, and residuals obtained from this process were used to train the LSTM model. The result of the hybrid model was compared with Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) as evaluation metrics. The hybrid of LSTM-ARIMA demonstrates superior evaluation performance than traditional ARIMA and LSTM models. The results also demonstrated that using the hybrid model of LSTM-ARIMA is more effective and reliable in predicting the maternal mortality rate. This can be applied to other cases of general time series forecasting, which improves the existing traditional models with modern approaches and machine learning concepts.
逐年被引趋势
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
计算机 / AIForecasting Techniques and Applications
Stock Market Forecasting Methods · Time Series Analysis and Forecasting