本文選題：特性曲線 + SVM　； 參考：《給水排水》2014年S1期

【摘要】：針對水泵揚程數(shù)據(jù)測量的有限性,并根據(jù)支持向量機能夠很好地處理小樣本事件能力的特點,提出了基于支持向量機的水泵數(shù)據(jù)挖掘方法。該方法旨在最有限的測量數(shù)據(jù)情況下挖掘出測量數(shù)據(jù)中間點的數(shù)據(jù),從而繪制出最接近真實的流量—揚程曲線圖。數(shù)據(jù)挖掘后形成的曲線圖與一般的樣本圖的趨勢相比較,支持向量機的數(shù)據(jù)挖掘模型好于神經(jīng)網(wǎng)絡(luò)法。
[Abstract]:In view of the limited measurement of pump head data and the ability of SVM to deal with small sample events, a water pump data mining method based on SVM is proposed. The purpose of this method is to mine the data of the middle point of the measurement data under the condition of the most limited measurement data, so as to draw the chart of the nearest real flow-head curve. Compared with the trend of the general sample diagram, the data mining model of support vector machine is better than the neural network method.
【作者單位】： 北京威派格科技發(fā)展有限公司;華北電力大學能源動力與機械工程學院;北京航空航天大學能源與動力工程學院;
【基金】：國家水體污染控制與治理科技重大專項(2012ZX07411-001)
【分類號】：TH38;TU991.35

【參考文獻】

相關(guān)期刊論文 前7條

1 孫相文;王美玉;惠t，

本文編號：1892695