Handling Missing Value and Outlier to Improve Model Performance for Predicting Logistic Case

Afif Amirullah, Riyanarto Sarno, Kelly Rossa Sungkono, Agus Tri Haryono, Abdullah Faqih Septiyanto, Sholiq

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The world of logistics is a world that urgently needs speed and accuracy in operational calculations. In order to support these operational needs, the logistics industry needs to keep up with the times by utilizing machine learning in its business processes. However, problems when utilizing machine learning in the industrial world include incomplete data and abnormal or outlier data caused by certain factors. This researcher will focus on improving machine learning models using KNN-Imputer to overcome incomplete data or missing values and Local Factor Outlier to overcome abnormal data. The results of the study show that when the extraction feature is utilized, the accuracy level of machine learning increases by 34.86% and decreases error by MAE 97.47%, MSE 99.72%, and RMSE 96.75%.

Original languageEnglish
Title of host publication2024 Beyond Technology Summit on Informatics International Conference, BTS-I2C 2024
EditorsFerry Wahyu Wibowo
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages77-82
Number of pages6
ISBN (Electronic)9798331508579
DOIs
Publication statusPublished - 2024
Externally publishedYes
Event2024 Beyond Technology Summit on Informatics International Conference, BTS-I2C 2024 - Jember, Indonesia
Duration: 19 Dec 2024 → …

Publication series

Name2024 Beyond Technology Summit on Informatics International Conference, BTS-I2C 2024

Conference

Conference2024 Beyond Technology Summit on Informatics International Conference, BTS-I2C 2024
Country/TerritoryIndonesia
CityJember
Period19/12/24 → …

Keywords

  • KNN-Imputer
  • Local Factor Outlier
  • Logistics
  • Machine Learning

Fingerprint

Dive into the research topics of 'Handling Missing Value and Outlier to Improve Model Performance for Predicting Logistic Case'. Together they form a unique fingerprint.

Cite this