Low Percentage Missing Imputation using KNN, NB and DT

The objective of this research is to test data imputation for Missing data over 7 cases. Different machine learning algorithms to impute the missing data were tested and evaluated: K-nearest Neighbor (KNN), Nave Bayes (NB) and Decision Tree (DT). Evaluation was done using t-test for the experiment with different configurations (i. e.5 %, 10 % missing). The result of the experiment shows that KNN has scored better results compared with Nave Bayes and Decision Tree. In conclusion, it is clear that machine learning algorithms can be used for missing data imputation. The implications of this research shows promising potentials for the utilization of KNN