PENGEMBANGAN MODEL UNTUK MEMONITOR JUMLAH SAMPAH DI KOTA YOGYAKARTA
Waste is residual human activities which become a national issue in Indonesia, including in Yogyakarta city. Our previous research developed Bayesian Network method to complete the waste volume prediction in Yogyakarta based on some external factors. The research came up with the new model to predic...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Theses and Dissertations NonPeerReviewed |
| Published: |
[Yogyakarta] : Universitas Gadjah Mada
2014
|
| Subjects: | |
| Online Access: | https://repository.ugm.ac.id/133864/ http://etd.ugm.ac.id/index.php?mod=penelitian_detail&sub=PenelitianDetail&act=view&typ=html&buku_id=74740 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Universitas Gadjah Mada |
| Summary: | Waste is residual human activities which become a national issue in Indonesia, including in Yogyakarta city. Our previous research developed Bayesian Network method to complete the waste volume prediction in Yogyakarta based on some external factors. The research came up with the new model to predict waste volume whether it is decreasing or increasing. However, the model has not been able to obtain the decreasing or increasing number of waste volume in Yogyakarta. Moreover, the accuracy of this model is considered too low because of missing value suspicion in the data used. Therefore, this research develops a waste volume monitoring model by considering historical data of waste volume in Yogyakarta (closed system) and also some external factors that may influence waste volume in Yogyakarta (open system).
In this research, Holt�s model as a closed system model is combined with Bayesian Network model as an open system model to develop the waste volume monitoring model. The first step to develop the combination model is to process the data containing the missing value. Then, the results of the missing value processing are used to develop Holt�s and Bayesian Network for each model. The result of each model will be combined to obtain the number of waste volume in Yogyakarta when it is predicted. As a comparison, other combination models are developed using combination forecasting method, such as Simple Average, Inverse MSE Weights, and Odd Matrix. These methods are based on the two best of time series forecasting method among Naive, Simple Averages, Moving Averages, Single Exponential Smoothing, and ARIMA.
The result of Holt�s and Bayesian Network combination model test shows that this model has accuracy value MAPE as 14%, MAD as 7,580 x 105, and MSE as 1,096 x 1012 in predicting and monitoring the waste volume in Yogyakarta. The accuracy is better than the accuracy got from the other combination models that combines Moving Averages with Single Exponential Smoothing. Moreover, Holt�s and Bayesian Network combination model has ability to represent the waste volume actual trend as 67%. This rate is calculated based on the suitability frequency between model result pattern with actual trend pattern of waste volu |
|---|