All Times ET
Virtual
State Space Model Using Kalman Filter to Forecast Mortality Time Series Data in Health Science (304145)
*Jae J Lee, State University of New York, New PaltzKeywords: Time series data, Kalman filter, State space model, Mortality in US
This presentation shows how to convert ARIMA models and Structural Time series models to state space form and shows how to set up Kalman filter to forecast the mortality time series data in US.
Time series data can be modeled using ARIMA for non-seasonal time series or multiplicative ARIMA for seasonal time series. Also, time series data can be modeled using Structural time series model. Both models will be converted to a state space form and be analyzed by Kalman filter.
This presentation will use a number of US mortality time series data from several sources. Each mortality data will be modeled by ARIMA and/or Structural time series model. Then they will be converted to a state space form and will be analyzed by Kalman filter.
This presentation is for audiences with some basic levels of knowledge in time series analysis.