Classify Event Participants in Universities and Industries Using Knowledge Discovery in Databases
Keywords:Classification, Classifier Accuracy Measurables Method, Data Mining, KDD, Naïve Bayes
One of the needs of event organizers is to be able to find out the profile of event participants based on the training data that has been created, find out the relationship between the selected attributes (namely, education, age, job, filed, business scale, workplace, online marketing) and the class that has been formed, and find the relationship between prospective event participants and their eligibility to attend the event. The knowledge discovery in database could be applied both in universities and industries to classify the participant’s attributes. In addition, the organizers need to be able to analyze and group event participants into predetermined classes so as to facilitate the stages of the selection process. This research will analyze the data to find new knowledge using the Knowledge Discovery in Databases (KDD) stage and data mining techniques, namely Classification. Classification technique works by grouping data based on training data and the value of the classification attribute. The Classification method predicts the class of a thing from a set of attributes that describe it, where the prediction process is carried out based on a database that already has a class label first. Classifier learning is a selection process that has a high level of accuracy. Naive Bayes technique is one of the simple classification techniques but can provide a very good level of accuracy. This research will classify event participants into three classes, namely Eligible, Considered and Unfit classes based on predetermined attributes in which the participants from nearby locations with sound knowledge are preferred. One of the results of this research is expected to assist event organizers in conducting the analysis process of the relationship between event participants and the feasibility of participating in the event from the data that has been obtained. Based on the results of the performance evaluation of the classification method using the Classifier Accuracy Measurables method, it shows that the accuracy of the predictions that have been made is 94%.