PREDICTION OF INCOMING ORDERS USING THE LONG SHORT-TERM MEMORY METHOD AT PT. XYZ
Currently the need for domestic packaging paper continues to increase, driven by the level of consumer awareness about sustainable packaging. PT XYZ is a local company engaged in the Corrugated Cardboard Box (KKG) industry. So far, the problems in fulfilling incoming orders every month are not optim...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Program Studi Teknik Informatika Universitas Trilogi
2021-06-01
|
| Series: | JISA (Jurnal Informatika dan Sains) |
| Subjects: | |
| Online Access: | https://trilogi.ac.id/journal/ks/index.php/JISA/article/view/902 |
| _version_ | 1798000187604467712 |
|---|---|
| author | Lukman Irawan Fauzi Fauzi Denny Andwiyan |
| author_facet | Lukman Irawan Fauzi Fauzi Denny Andwiyan |
| author_sort | Lukman Irawan |
| collection | DOAJ |
| description | Currently the need for domestic packaging paper continues to increase, driven by the level of consumer awareness about sustainable packaging. PT XYZ is a local company engaged in the Corrugated Cardboard Box (KKG) industry. So far, the problems in fulfilling incoming orders every month are not optimal with an average of about 30% inaccuracy. This is because the orders that enter cannot be predicted. As an effort to win market competition in packaging paper, PT. XYZ must improve the fulfillment of incoming orders by predicting incoming orders using the Long Short-Term Memory (LSTM) method. The aim of this research is to provide a predictive model for incoming orders in accordance with the needs of order fulfillment to be applied to production planning. So that order fulfillment can be on time. The method used in predicting incoming orders is the Long Short-Term Memory (LSTM) method using weighting evaluations with the lowest Root Mean Squared Error (RMSE) and Augmented Dickey-Fuller test (ADF). The test results of the LSTM method with parameter sizes of Batch: 1 Epochs: 5000 Neurons: 1 show that the RMSE for MDM products is 8.767582 and 0.287924, LNR products are 10.623984 and 0.466621, WTP products are 1.636849 and 0.361515 lower than the size of the fit parameters for other LSTM models, and the ADF Statistic value for MDM products -6.137597, LNR -6.753697, WTP -4.872927 |
| first_indexed | 2024-04-11T11:16:11Z |
| format | Article |
| id | doaj.art-6627bbaa27a64efd939ffbdd6149600e |
| institution | Directory Open Access Journal |
| issn | 2776-3234 2614-8404 |
| language | English |
| last_indexed | 2024-04-11T11:16:11Z |
| publishDate | 2021-06-01 |
| publisher | Program Studi Teknik Informatika Universitas Trilogi |
| record_format | Article |
| series | JISA (Jurnal Informatika dan Sains) |
| spelling | doaj.art-6627bbaa27a64efd939ffbdd6149600e2022-12-22T04:27:14ZengProgram Studi Teknik Informatika Universitas TrilogiJISA (Jurnal Informatika dan Sains)2776-32342614-84042021-06-0141808910.31326/jisa.v4i1.902539PREDICTION OF INCOMING ORDERS USING THE LONG SHORT-TERM MEMORY METHOD AT PT. XYZLukman Irawan0Fauzi Fauzi1Denny Andwiyan2Budi Luhur UniversityBudi Luhur UniversityRaharja UniversityCurrently the need for domestic packaging paper continues to increase, driven by the level of consumer awareness about sustainable packaging. PT XYZ is a local company engaged in the Corrugated Cardboard Box (KKG) industry. So far, the problems in fulfilling incoming orders every month are not optimal with an average of about 30% inaccuracy. This is because the orders that enter cannot be predicted. As an effort to win market competition in packaging paper, PT. XYZ must improve the fulfillment of incoming orders by predicting incoming orders using the Long Short-Term Memory (LSTM) method. The aim of this research is to provide a predictive model for incoming orders in accordance with the needs of order fulfillment to be applied to production planning. So that order fulfillment can be on time. The method used in predicting incoming orders is the Long Short-Term Memory (LSTM) method using weighting evaluations with the lowest Root Mean Squared Error (RMSE) and Augmented Dickey-Fuller test (ADF). The test results of the LSTM method with parameter sizes of Batch: 1 Epochs: 5000 Neurons: 1 show that the RMSE for MDM products is 8.767582 and 0.287924, LNR products are 10.623984 and 0.466621, WTP products are 1.636849 and 0.361515 lower than the size of the fit parameters for other LSTM models, and the ADF Statistic value for MDM products -6.137597, LNR -6.753697, WTP -4.872927https://trilogi.ac.id/journal/ks/index.php/JISA/article/view/902predictionplanninglstmtime series |
| spellingShingle | Lukman Irawan Fauzi Fauzi Denny Andwiyan PREDICTION OF INCOMING ORDERS USING THE LONG SHORT-TERM MEMORY METHOD AT PT. XYZ JISA (Jurnal Informatika dan Sains) prediction planning lstm time series |
| title | PREDICTION OF INCOMING ORDERS USING THE LONG SHORT-TERM MEMORY METHOD AT PT. XYZ |
| title_full | PREDICTION OF INCOMING ORDERS USING THE LONG SHORT-TERM MEMORY METHOD AT PT. XYZ |
| title_fullStr | PREDICTION OF INCOMING ORDERS USING THE LONG SHORT-TERM MEMORY METHOD AT PT. XYZ |
| title_full_unstemmed | PREDICTION OF INCOMING ORDERS USING THE LONG SHORT-TERM MEMORY METHOD AT PT. XYZ |
| title_short | PREDICTION OF INCOMING ORDERS USING THE LONG SHORT-TERM MEMORY METHOD AT PT. XYZ |
| title_sort | prediction of incoming orders using the long short term memory method at pt xyz |
| topic | prediction planning lstm time series |
| url | https://trilogi.ac.id/journal/ks/index.php/JISA/article/view/902 |
| work_keys_str_mv | AT lukmanirawan predictionofincomingordersusingthelongshorttermmemorymethodatptxyz AT fauzifauzi predictionofincomingordersusingthelongshorttermmemorymethodatptxyz AT dennyandwiyan predictionofincomingordersusingthelongshorttermmemorymethodatptxyz |