<![CDATA[Abstrak - Penelitian ini mengembangkan sistem prediksi harga mobil bekas menggunakan Algoritma Random Forest dengan menggunakan dataset Car Dekho Used Car Prices yang diperoleh dari platform Kaggle. Himpunan data tersebut mencakup 15411 row data mobil dengan 14 atribut, dan memiliki total data sebanyak 46.322. Proses penelitian meliputi data collecting, exploratory data analysis, data preprocessing (pengecekan missing values dan encoding variabel kategorikal), pembagian data menggunakan train-test split pembagian ini dilakukan dengan perbandingan data Training set (80%) dan data Testing set (20%). Hasil analisis menunjukkan bahwa fitur km_driven, vehicle_age, max_power, dan mileage memiliki pengaruh paling signifikan terhadap prediksi harga. Model Random Forest memberikan performa prediksi yang baik dengan nilai MAE sebesar Rupee130.309, RMSE sebesar Rupee212.103, dan R² sebesar 0,9505, yang berarti model mampu menjelaskan 95,05% variasi harga jual mobil bekas. Model kemudian diimplementasikan dalam aplikasi berbasis Streamlit yang memungkinkan pengguna memasukkan spesifikasi kendaraan dan memperoleh estimasi harga secara interaktif. Sistem prediksi ini dapat membantu konsumen dan pelaku bisnis dalam menentukan harga mobil bekas secara lebih akurat dan informatif.Kata kunci : Prediksi Harga Mobil Bekas; Random Forest; Machine Learning; Streamlit; Abstract - This research developed a used car price prediction system using the Random Forest algorithm with the Car Dekho Used Car Prices dataset obtained from the Kaggle platform. The dataset includes 15,411 rows of car data with 14 attributes, for a total of 46,322 data points. The research process includes data collecting, exploratory data analysis, data preprocessing (checking missing values and encoding categorical variables), and data division using train-test split. This division is carried out with a comparison of the training set (80%) and testing set (20%) data. The analysis results show that the features km_driven, vehicle_age, max_power, and mileage have the most significant influence on price prediction. The Random Forest model provided good prediction performance with an MAE value of IDR 130,309, an RMSE of IDR 212,103, and an R² of 0.9505, which means that the model was able to explain 95.05% of the variation in used car prices. The model was then implemented in a Streamlit-based application that allows users to enter vehicle specifications and obtain price estimates interactively. This prediction system can help consumers and businesses determine used car prices more accurately and informatively.Keywords: Used Car Price Prediction; Random Forest; Machine Learning; Streamlit;]]>
