基于改進(jìn)k-means算法的海量智能用電數(shù)據(jù)分析
發(fā)布時(shí)間：2019-01-18 15:48
【摘要】：針對(duì)智能用電數(shù)據(jù)挖掘面臨數(shù)據(jù)量大、挖掘效率低等難題,進(jìn)行Map-Reduce模型下基于改進(jìn)k-means的海量用電數(shù)據(jù)分析研究。以家庭用戶為例,建立了家庭用戶用電信息的家庭用戶號(hào)、房屋面積、家庭成員數(shù)、每天用電量、峰谷電量、家用電器數(shù)等的數(shù)據(jù)維度模型,利用k-means算法簡單、收斂速度快的優(yōu)勢,克服其容易陷入局部最優(yōu)解的缺陷,綜合考慮初始聚類中心的選擇及聚類個(gè)數(shù)的選取2個(gè)因素,以數(shù)據(jù)對(duì)象密度的大小作為初始聚類中心的選取標(biāo)準(zhǔn),將簇間距離及簇內(nèi)對(duì)象的分散程度作為聚類數(shù)目選擇的重要參考,對(duì)k-means算法進(jìn)行改進(jìn);為提高數(shù)據(jù)處理效率,進(jìn)行Map-Reduce處理模型下的海量家庭用戶用電數(shù)據(jù)的并行挖掘。通過在Hadoop集群上進(jìn)行實(shí)驗(yàn),結(jié)果證明提出的算法運(yùn)行穩(wěn)定、高效、可行,且具有良好的加速比。
[Abstract]:Aiming at the problems of large data volume, low mining efficiency and the like, the intelligent power consumption data mining is carried out under the Map-Reduce model, and the mass utilization data analysis research based on the modified k-means is carried out under the Map-Reduce model. a household user number, a house area, a family member number, a daily power consumption, a peak-to-valley electric quantity, a household appliance number and the like are established by using a household user as an example, a data dimension model such as a peak-to-valley electric quantity, a household appliance number and the like is established, the defect that the local optimal solution is easy to fall into a local optimal solution is overcome, the selection of the initial poly-type center and the selection of the number of the poly-classes are comprehensively considered, the size of the data object density is taken as the selection criterion of the initial clustering center, In order to improve the data processing efficiency, the parallel mining of the data of mass home users under the Map-Reduce processing model is carried out to improve the data processing efficiency. The results show that the proposed algorithm is stable, efficient and feasible and has a good speedup ratio by doing the experiments on the Hadoop cluster.
【作者單位】： 重慶市電力公司;
【分類號(hào)】：TM769;TP311.13
[Abstract]:Aiming at the problems of large data volume, low mining efficiency and the like, the intelligent power consumption data mining is carried out under the Map-Reduce model, and the mass utilization data analysis research based on the modified k-means is carried out under the Map-Reduce model. a household user number, a house area, a family member number, a daily power consumption, a peak-to-valley electric quantity, a household appliance number and the like are established by using a household user as an example, a data dimension model such as a peak-to-valley electric quantity, a household appliance number and the like is established, the defect that the local optimal solution is easy to fall into a local optimal solution is overcome, the selection of the initial poly-type center and the selection of the number of the poly-classes are comprehensively considered, the size of the data object density is taken as the selection criterion of the initial clustering center, In order to improve the data processing efficiency, the parallel mining of the data of mass home users under the Map-Reduce processing model is carried out to improve the data processing efficiency. The results show that the proposed algorithm is stable, efficient and feasible and has a good speedup ratio by doing the experiments on the Hadoop cluster.
【作者單位】： 重慶市電力公司;
【分類號(hào)】：TM769;TP311.13
【參考文獻(xiàn)】
相關(guān)期刊論文 前10條
1 李智勇;吳晶瑩;吳為麟;宋保明;;基于自組織映射神經(jīng)網(wǎng)絡(luò)的電力用戶負(fù)荷曲線聚類[J];電力系統(tǒng)自動(dòng)化;2008年15期
2 肖世杰;;構(gòu)建中國智能電網(wǎng)技術(shù)思考[J];電力系統(tǒng)自動(dòng)化;2009年09期
3 劉友波;劉俊勇;趙巖;李磊;胥威汀;姚s，
本文編號(hào)：2410866
本文鏈接：http://sikaile.net/kejilunwen/dianlilw/2410866.html
最近更新
教材專著
