Implementation The Convolutional Neural Network Method For Classification The Draw-A-Person Test

Psychotest or psychological tests at this time are often applied in the process of selection of human resources aimed at measuring the potential of intelligence, recognizing personality, predicting work performance, mapping potential, and level of productivity. The Draw-A-Person test has been long applied to measure personality and to know the individual's creative experience. This test is widely used by psychologist institution in Psychotest because the implementation of test is quite simple that only use a pencil as well as paper. In practice, a psychologist takes quite a long time to assess the result of the Draw-A-Person test. To accelerated the required time and facilitate the work of a psychologist, a model is needed to recognize and classify the results of a Draw-A-Person test. This model is able to recognize and study the Draw-A-Person test result based on the head-size drawings on paper. Deep learning with a convolutional neural network method is applied to recognize and study the Draw-A-Person test result. To improve the usability of CNN method, the data is in the form of a digital image. The data is collected using a smartphone camera and labeled in Microsoft Excel one by one according to the criteria on the image. Data that has been labeled will be used to train the model. The trained model will be tested for new data. In this research, the data train achieves 99.48% accuracy and 1.74% loss. In the new data, the model achieved 66.7% accuracy.