Identification of Drug Dependence Types Using Case-Based Reasoning Expert System with K-Nearest Neighbour

Drugs are a serious concern in developing countries, including Indonesia. Drug users, who are increasing every year, are a global problem that must be considered, not only by the government but also by the community. This increase is partly due to the ignorance of drug users about the type of drug dependence on themselves, making it difficult to cure themselves. Identification of dependence is very necessary. Therefore, this study uses the Case-Based Reasoning (CBR) method based on an expert system with K-Nearest Neighbor (KNN) to classify types of drug dependence. Types of drug dependence are based on the class of substances, namely Stimulants, Opioids (narcotics), Depressants, and Hallucinogens. The research data uses medical record data of drug users from 2018 to 2020. The results of the CBR method calculation with K-NN using 72 old case data and 30 new case data with confusion matrix testing obtained good accuracy results based on accuracy values of $\mathbf{8 1 \%} \%, \mathbf{7 6 \%}$ for precision, and $\mathbf{1 0 0 \%}$ for recall. From the test results, it can be seen that the CBR expert system with K-NN can identify types of drug dependence with a high level of accuracy.