Identification of Indonesian Authors Using Deep Neural Networks

Author Name Disambiguation (AND) is a problem that occurs when set of publications contains ambiguous names authors, i.e. the same author may appear with different (synonyms) in other published papers, or (authors) who be have name (homonym). In this final project, we will design model Deep Neural Network (DNN) classifier. The dataset used project uses primary data sourced from Scopus website. This research focuses on integrating Indonesian authors. Parameters accuracy, sensitivity and precision are standard benchmarks to determine performance method solve AND problems. best DNN classification achieves 99.9936% Accuracy, 93.1433% Sensitivity, 94.3733% Precision. Then for highest measurement, case Non Synonym-Homonym (SH) has 99.9967% 96.7388% 97.5102%