Waste generation is one of the main problems in the waste problem. Factors that influence the amount of waste generated are population and weather. The volume of waste in landfills continues to increase along with population growth. Therefore, it needs to be handled through a waste volume prediction approach. Many studies have been carried out on predicting waste in a city, but predictions have yet to be carried out at a more detailed level. This study proposes a time-series forecasting approach to accurately predict waste, utilizing a one-layer LSTM network and multivariate time-series data. A waste volume prediction model was developed for each urban village in Magelang City, considering the influence of population and weather factors with time lag. The data that has been preprocessed is subjected to correlation analysis using Pearson correlation and statistical significance tests to produce significant variables as modeling input. The research showed that not all factors influencing waste volume predictions are significant in all urban village. The LSTM model is optimal with an RMSE value of 0.0786 for the Tidar Selatan urban village and an value of 0.4462 for the Rejowinangun Utara urban village. All factors used in this research are significant for the two urban village. However, significant urban village with all the factors that influence them produce RMSE and values that could be more optimal.

Original languageEnglish
Title of host publication2023 8th International Conference on Informatics and Computing, ICIC 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350342604
Publication statusPublished - 2023
Event8th International Conference on Informatics and Computing, ICIC 2023 - Hybrid, Malang, Indonesia
Duration: 8 Dec 20239 Dec 2023

Publication series

Name2023 8th International Conference on Informatics and Computing, ICIC 2023


Conference8th International Conference on Informatics and Computing, ICIC 2023
CityHybrid, Malang


  • LSTM
  • multivariate
  • significant variable
  • time lag
  • waste prediction


Dive into the research topics of 'The Use of LSTM Model with Lagged Daily Inputs for Waste Disposal Prediction'. Together they form a unique fingerprint.

Cite this