Implementation of the AI-Based Xgboost Algorithm in a Used Car Price Prediction Web System

Used car pricing in Indonesia still often faces major obstacles in the form of information asymmetry, subjectivity of appraisers' assessments, and the complex relationship between vehicle attributes that affect market selling prices. This study aims to implement an artificial intelligence-based Extreme Gradient Boosting (XGBoost) algorithm to predict used car prices precisely and integrate it into an interactive web-based system ecosystem. The research methodology refers to the Cross-Industry Standard Process for Data Mining (CRISP-DM) industry standard framework which includes six main stages, namely business understanding, data understanding, data preparation, modeling, evaluation, and deployment. Data collection was carried out through web scraping techniques on the Carmudi Indonesia digital platform, which successfully captured 5,106 rows of raw data. The data preparation stage includes the process of data cleaning, handling invalid data, removing outliers using the Interquartile Range (IQR) method, forming vehicle age variables, and applying Ordinal Encoding to all categorical attributes, resulting in 4,843 clean data. The data was then partitioned proportionally into 80% training data (3,874 data) and 20% testing data (969 data). Modeling used XGBoost Regressor involving nine main predictor variables. Model performance evaluation was carried out using the Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination R2. The test results showed a very high level of performance with an R2 value of 93.13%, MAE of Rp41,090,000, RMSE of Rp91,370,000, and MAPE of 12.58%. Furthermore, the trained model was serialized and integrated into a web application based on the Flask Python framework so that the system is able to generate real-time price estimates. This implementation is expected to function as an objective and transparent decision support instrument for the wider community.

Authors

Institutions

Publication Details

Journal
SainsTech Innovation Journal
Published
2026-11-13
DOI
https://doi.org/10.37824/sij.v9i2.2026.1521
Primary Topic
Data Mining and Machine Learning Applications
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Implementation of the AI-Based Xgboost Algorithm in a Used Car Price Prediction Web System

Zaid Amin, Kuncoro Adimas
SainsTech Innovation Journal
Data Mining and Machine Learning Applications
article

Implementation of the AI-Based Xgboost Algorithm in a Used Car Price Prediction Web System

Zaid Amin, Kuncoro Adimas
article en

Abstract

Used car pricing in Indonesia still often faces major obstacles in the form of information asymmetry, subjectivity of appraisers' assessments, and the complex relationship between vehicle attributes that affect market selling prices. This study aims to implement an artificial intelligence-based Extreme Gradient Boosting (XGBoost) algorithm to predict used car prices precisely and integrate it into an interactive web-based system ecosystem. The research methodology refers to the Cross-Industry Standard Process for Data Mining (CRISP-DM) industry standard framework which includes six main stages, namely business understanding, data understanding, data preparation, modeling, evaluation, and deployment. Data collection was carried out through web scraping techniques on the Carmudi Indonesia digital platform, which successfully captured 5,106 rows of raw data. The data preparation stage includes the process of data cleaning, handling invalid data, removing outliers using the Interquartile Range (IQR) method, forming vehicle age variables, and applying Ordinal Encoding to all categorical attributes, resulting in 4,843 clean data. The data was then partitioned proportionally into 80% training data (3,874 data) and 20% testing data (969 data). Modeling used XGBoost Regressor involving nine main predictor variables. Model performance evaluation was carried out using the Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination R2. The test results showed a very high level of performance with an R2 value of 93.13%, MAE of Rp41,090,000, RMSE of Rp91,370,000, and MAPE of 12.58%. Furthermore, the trained model was serialized and integrated into a web application based on the Flask Python framework so that the system is able to generate real-time price estimates. This implementation is expected to function as an objective and transparent decision support instrument for the wider community.

SainsTech Innovation Journal
Universitas Bina Darma (ID)
Data Mining and Machine Learning Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon