The normalizing method of different batch samples in illustrated with the example of high - voltage switch 以高壓開關(guān)管為例,介紹了不同批樣本的歸一化處理方法。
In order to make it sure that the data is in same quantity rank , the author adopts normalized method to treat the data imported and exported in advance while training bp net 在進(jìn)行網(wǎng)絡(luò)訓(xùn)練時(shí),為了保證數(shù)據(jù)為同一個(gè)數(shù)量級(jí),采用了歸一化處理方法,對(duì)神經(jīng)網(wǎng)絡(luò)的輸入和輸出數(shù)據(jù)進(jìn)行一定的預(yù)處理。
This paper will discuss how we can revise the offset by curve matching , the algorithm of the neural network and the normalized method of the polynomial match so that the sensor price will fall and its accuracy will be improved 本論文應(yīng)用曲線擬合方法,神經(jīng)網(wǎng)絡(luò)算法和多項(xiàng)式擬合的規(guī)范化方法校正零點(diǎn),降低成本且精確度提高。