Aplikasi Metode Regresi Linier Berganda dan Jaringan Syaraf Tiruan Untuk Memrediksi Kinerja Usaha Mikro dan Kecil

Abstract

An increasing number of population each year always increase the number of labor force and job demand. This is the main task of the government as the main role and societies to actively participate in the development. Establishment and development of micro and small enterprises become one solution. Ability of societes to establish and expand these enterprises must obtain government support. One of the forms of government support is training for their managers/owners in order to increase their performance. The effect of training on the performance can be analyzed by multiple regression, classification and decision trees, neural networks, and so on. This study aims to compare the artificial neural network and multiple linear regression to predict the performance of these enterprises by considering implementation of the training and implementation of the training outcomes with sample of 60 units. The comparison of the prediction accuracy rate of these methods use mean absolute error (MAE), mean absolute percentage error (MAPE), and mean square error (MSE). The results of this study shows that multiple linear regression method produces higher accuracy rate for predicting the performance of micro and small enterprises than artificial neural network method