IJRR

International Journal of Research and Review

| Home | Current Issue | Archive | Instructions to Authors | Journals |

Research Paper

Year: 2020 | Month: March | Volume: 7 | Issue: 3 | Pages: 151-160

KNN Imputation Missing Value For Predictor App Rating On Google Play Using Random Forest Method

Abdul Khaliq1, Pahala Sirait 2, Andri2

1Postgraduate Student at STMIK Mikroskil, Medan, Indonesia, 20212
2Lecturer at STMIK Mikroskil, Medan, Indonesia, 20212

Corresponding Author: Abdul Khaliq

ABSTRACT

Developers and application users are key to the market's impact on application development. In the development of application, developers need to accurately predict applications in the market, accurate prediction results are crucial in showing the rating of the user affects the success of an application. In data retrieval, there is missing data. Lost Data is done by the process of missing value Imputasi using KNN Imputation. Predictions will be done using the random forest algorithm as a method used to predict app ratings. The combination of the KNN method for the first imputation of using a random forest algorithm is better than without imputation. It can be seen from using a random forest algorithm with an average of 91,4465% accuracy results, the result is better than the prediction without the imputation of the missing value with an accuracy result of 75,8465%.

Keywords: Imputation Missing Value, App Rating, Prediction, KNN Imputation, Random Forest

[PDF Full Text]