@thesis{thesis,
		author={Dimas  Purwitasari},
		title ={Perbaikan Nilai Akurasi Metode Backpropagation Menggunakan Metode Fuzzy Tsukamoto (Studi Data : Kasus Kumulatif Covid-19 Di Indonesia)},
		year={2021},
		url={https://repository.ittelkom-pwt.ac.id/6701/},
		abstract={The backpropagation method is one of the most frequently used methods, but this method also has weaknesses, namely instability and poor convergence speed, this is due to being trapped in an artificial neural network at a minimum local level. Based on previous research to predict COVID-19 cases with a combination of Backpropagation and Fuzzy Tsukamoto methods by obtaining a Mean Square Error (MSE) value for normalized data of 1,632337, however, in this study, researchers only used one experiment on the Backpropagation network architecture. In this study, researchers used the fuzzy tsukamoto method to deal with weaknesses in the backpropagation method. Tsukamoto fuzzy method is used to generate learning speed and momentum and perform 10 (ten) experiments on backpropagation network architecture. The results of this study show the average accuracy of MSE is 0,16278. By using the lerning rate and momentum parameters, the results of testing on the fuzzy tsukamoto method are 0,0025 and 0,25, as well as the other best parameters, among others, windowing 9, hidden layer 5 and epoch 1000. Keywords : Accuracy, Backpropagation, COVID-19, Fuzzy Tsukamoto, Mean Square Error (MSE).}
		}