COMMAX 2020 UiTM Kampus Kuala Terengganu

Social anxiety disorder is the intense fear to social situations which involves interaction with stranger or unfamiliar person. There are three levels of social anxiety which are mild anxiety, moderate anxiety and severe anxiety. Social anxiety disorder cannot be diagnosed by physical exam or medical test. The healthcare provider will diagnose social phobia from a description of the symptoms. The main objective of this study is to develop a prototype which is Social Anxiety Disorder Diagnostic using K-Nearest Neighbor algorithm to diagnose the social anxiety disorder. The prototype gives an output of the level of social anxiety disorder which are mild, moderate and severe. K-Nearest Neighbor is one of the most widely used machine learning algorithm in healthcare problem. The prototype is build based on the research framework. There are five phases in the research framework which is theoretical study, data collection and preparation, algorithm design, implementation and testing and evaluation. Theoretical study phase is done by reading journal and article paper of social anxiety disorder, machine learning and K-Nearest Neighbor. Data collection and preparation phase is done by using secondary data acquisition. The data is collected from the internet. In algorithm design phase, data pre-processing and K-Nearest Neighbor classification is done. Data pre-processing is done by extract important data that is needed for the classification. The K-Nearest Neighbor classification is done by complete the step in the classification. The steps are initialized value of K, compute distance using Euclidean distance, sort the distance, take K-Nearest Neighbor and apply simple majority. Implementation phase is done by writing the program using java and correcting and debugging logical and syntax errors. Testing and evaluation phase are done by testing the K-Nearest Neighbor classifier using 10-fold cross validation. The evaluation is done for this project using 10-fold cross validation with the accuracy of 0.8406.