Aprendizado incremental com interação para classificação de personas

Abstract

The use of personas to identify audiences help in decision making and service delivery. The company analyzed in this monograph conducted an internal research to determine their personas within employee profles and developed a survey system to classify them. The purpose of this classifcation is to improve the distribution of hardware to employees, ofering ideal options for each profle. With the progressive growth of employee mapping, the survey rule system may become obsolete and, therefore, the objective of this monograph is to analyze and propose an incremental model for classifying the company’s employees, bringing the possibility of replacing the survey system in the future. The tecnhnique used to develop the model was the incremental learning method for data streams, where the use of Adaptive Random Forests (ARF) and Support Vector Machines (SVM) classifers was presented. In addition, the proposed model is adapted to receive feedback on the result presented by the system administrator, causing the feedback to be incorporated into the learning process.