ANALYSIS OF GENERALIZED SPACE TIME AUTOREGRESSIVE MODEL WITH BAYESIAN NETWORK AND MINIMUM SPANNING TREE APPROACH OF WEIGHT MATRIX (CASE STUDY ON POSITIVE CASE INCREASE OF COVID-19 IN JAVA ISLAND)
The ongoing global Coronavirus 2019 (COVID-19) pandemic poses a major threat. The spread of the COVID-19 virus is likely to occur from one location to another due to the mobility of people moving from one location to another. Many efforts and policies have been made by each country to slow the sprea...
Saved in:
| Main Author: | |
|---|---|
| Format: | Theses |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/54950 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | The ongoing global Coronavirus 2019 (COVID-19) pandemic poses a major threat. The spread of the COVID-19 virus is likely to occur from one location to another due to the mobility of people moving from one location to another. Many efforts and policies have been made by each country to slow the spread of the COVID-19 outbreak. The imposition of lockdowns and large-scale social restrictions / Social Distancing has been widely used to limit the transmission of this virus at the community and provincial levels. The two policies have proven effective in reducing the spread of the COVID-19 virus. The GSTAR model was applied to model the increase in COVID-19 cases per day in six provinces in Java Island. Data on the increase in COVID-19 cases per day were recorded simultaneously in six locations, namely in the Provinces of Banten, Jakarta, West Java, Central Java, Yogyakarta Special Region, and East Java. In this study, another approach was used in constructing the weight matrix required to build the GSTAR model, namely the Bayesian Network model and the Minimum Spanning Tree model. These two models will help build and represent the relationship between one location and another. By using the Bayesian network model, a probability graph (directed) model was created which states the causal relationship between six provinces in Java Island. Then by using the MST, a topological (undirected) network model was created to show the correlation, centrality, and relationship of the increase in positive cases of COVID-19 between provinces in Java Island. Based on the Mean Absolute Percentage Error, it is found that the weight matrix generated by the two models gives the same results in estimation (in-sample) and forecasting (out-sample) with the ordinary matrix. |
|---|