Development-and-research-of-neural-network-classifier-of-frequency-manipulated-signals

Logistic Regression, MLP, CNN The task is to classify sinusoidal signals with white noise superimposed on them. The distribution of noise is uniform f(x) = 1/(b-a), a=0, b=0..50 As classifiers were used Logistic Regression, MLP, CNN Dataset: 10000 training data, 1000 test data Quality of work of classifiers was estimated by means of a metric F1 To compare three classificators were found quantilies of F1(25%, 50%, 75%)