摘要或動機

Now Artificial Neural Networks using on the basic math is fewer. This paper is to suggest the Linear equation of the basic math using the BP Artificial Neural Networks. The BP Neural Networks have power ability for learning and can approximate any function, and regularity can be found to solve the linear equation. A good sample is one of the important elements for learning of Artificial Neural Networks. Generally, the samples are a lot of amount for the resolution of Linear equation. This paper is to use the principle of two points decide one line for the samples. The experiment shows that this method curtails many samples. Furthermore we also use Artificial Neural Networks to solve the problem of point-slope form. The experiment result is very satisfactory, and it offers some idea for the basic math using Artificial Neural Networks.目前人工神經網路較少用於基礎數學方面的求解，本文針對基礎數學直線方程式提出BP 人工神經網路應用於求解直線方程式，運用其很強的學習能力、(輸入向量和其對應的目標向量來訓練網路、逼近函數)，尋求規律來求解直線方程式；而良好的樣本是人工神經網路學習的重要條件之一，一般解決直線方程式需要大量樣本，本文利用二點決定一直線的原理來解決樣本問題，實驗結果顯示，這一方法成功的縮短了可觀的學習樣本，此外我們也運用BP 人工神經網路來求解點斜式的直線方程式問題，實驗結果是可行的，並且為人工神經網路用於基礎數學提供了一些思考方向。