No image available for this title

Text

Application of the LSTM Algorithm in Predicting Urea Fertilizer Production at IIB Plant PT. Pupuk Sriwidjaja Palembang



PT. Pupuk Sriwidjaja Palembang is a pioneer of fertilizer manufacturers in Indonesia. One of the plants at PT. Pupuk Sriwidjaja Palembang, namely the IIB urea plant, has been operating normally since 2017, thereby the data of production results has been collected for more than five years (time series data). The collected data can be used to make predictions of future production using the LSTM (Long Short Term Memory) model. LSTM is an artificial neural network architecture that is suitable for processing sequential data. The research objective to be achieved is to produce a production prediction model using LSTM modeling. Data collected over five years was divided into training data and testing data through data composition trials. The LSTM model training was carried out with a training data composition of 70% of the total data, batch size 64, and epoch 200. Then testing was carried out with data testing as much as 30% of the total data using RMSE and MAPE as model quality assessment parameters. Based on test results, the LSTM model is able to predict production with an RMSE of 11.08 and a MAPE of 6.39%.


Availability

No copy data


Detail Information

Series Title
-
Call Number
-
Publisher Jurnal Sains Komputer & Informatika (J-SAKTI) : Indonesia.,
Collation
005.3
Language
English
ISBN/ISSN
2548-9771
Classification
NONE
Content Type
-
Media Type
-
Carrier Type
-
Edition
-
Subject(s)
Specific Detail Info
-
Statement of Responsibility

Other Information

Accreditation
-

Other version/related

No other version available


File Attachment

  • Application of the LSTM Algorithm in Predicting Urea Fertilizer Production at IIB Plant PT. Pupuk Sriwidjaja Palembang
    PT. Pupuk Sriwidjaja Palembang is a pioneer of fertilizer manufacturers in Indonesia. One of the plants at PT. Pupuk Sriwidjaja Palembang, namely the IIB urea plant, has been operating normally since 2017, thereby the data of production results has been collected for more than five years (time series data). The collected data can be used to make predictions of future production using the LSTM (Long Short Term Memory) model. LSTM is an artificial neural network architecture that is suitable for processing sequential data. The research objective to be achieved is to produce a production prediction model using LSTM modeling. Data collected over five years was divided into training data and testing data through data composition trials. The LSTM model training was carried out with a training data composition of 70% of the total data, batch size 64, and epoch 200. Then testing was carried out with data testing as much as 30% of the total data using RMSE and MAPE as model quality assessment parameters. Based on test results, the LSTM model is able to predict production with an RMSE of 11.08 and a MAPE of 6.39%.


Information


Web Online Public Access Catalog - Use the search options to find documents quickly