【摘要】：為了處理大量分布式存儲的農(nóng)業(yè)環(huán)境數(shù)據(jù),實(shí)現(xiàn)農(nóng)業(yè)設(shè)施智能控制,基于內(nèi)存計(jì)算框架Spark提出一種并行化的Dirichlet過程混合模型聚類方法,對農(nóng)業(yè)環(huán)境及設(shè)施數(shù)據(jù)進(jìn)行訓(xùn)練得到預(yù)測模型,執(zhí)行對溫室大棚天窗開度的預(yù)測任務(wù)。通過對比實(shí)驗(yàn)驗(yàn)證了模型預(yù)測的可行性,對預(yù)測正確率進(jìn)行統(tǒng)計(jì),并測試了所提出的并行化聚類的執(zhí)行效率。實(shí)驗(yàn)結(jié)果表明,提出的方法具有較高的執(zhí)行效率及準(zhǔn)確性。
[Abstract]:In order to process a large amount of distributed storage of agricultural environment data and realize intelligent control of agricultural facilities, a parallel hybrid model clustering method of Dirichlet process is proposed based on memory computing framework Spark. The forecast model of agricultural environment and facility data was trained, and the forecast task of greenhouse skylight opening was carried out. Through the comparison experiment, the feasibility of the model prediction is verified, the accuracy of the prediction is counted, and the execution efficiency of the proposed parallel clustering is tested. The experimental results show that the proposed method has high efficiency and accuracy.
【作者單位】： 上海大學(xué)機(jī)電工程與自動化學(xué)院;上海市電站自動化技術(shù)重點(diǎn)實(shí)驗(yàn)室;
【基金】：上海市科委重點(diǎn)項(xiàng)目(14DZ1206302)
【分類號】：S625;TP311.13
，

本文編號：2445598