Smart forecast of Algeria's unemployment rates during 199-2020

الملخص EN

The current study has aimed to apply a mathematical approach of artificial intelligence, which is represented by non-linear autoregressive artificial neural network model (NNAR) to forecast Algeria's monthly unemployment rates during Dec 1991-Dec 2020, using various algorithms in the training process.

by comparing the results of proposed neural networks, it has been shown that the neural network model: NAR-LM of (4-1-20-1), which based on the back-propagation algorithm, has better performance than the Bayesian regulation neural network model and gradual training algorithm as well, where forecast error has reached a value of 3, 56*10-6.

also, the generated series by neural networks: NAR-LM (4.

1.

20.

1) and NAR-BR (4.

1.

20.

1) emulate well the original series compared to NAR-SCG (2.1.10.1).