Perbandingan Algoritma Adaline Berdasarkan Pola Input Data Dan Aktivasi Output Untuk Prediksi Data

Donni Nasution (Universitas Prima Indonesia, Medan, Indonesia)
Darmeli Nasution (Universitas Pembangunan Panca Budi, Medan, Indonesia)
S Solikhun (STIKOM Tunas Bangsa, Pematang Siantar, Indonesia)

Brahmana : Jurnal Penerapan Kecerdasan Buatan
Vol 4, No 1 (2022): Edisi Desember

Adaline is a single-layer supervised learning algorithm where the input layer is directly related to the output layer. Adaline learning uses the delta rule, which adjusts the weights to reduce the difference between network inputs to the desired output and output units. The main problem of this study is to find an alternative to the Adaline algorithm for predicting stroke with seven symptom attributes. This study seeks the best Adaline algorithm performance by comparing four forms of input and output activation patterns. The test results show the results of the same accuracy that is equal to 100%; the same epoch, namely one epoch, and the average weight change is different. The Adaline algorithm can predict stroke well with 100% accuracy.