溫榮坤
摘 要: 為了提高大數(shù)據(jù)環(huán)境下關(guān)聯(lián)數(shù)據(jù)挖掘的效率和精度,提出基于分?jǐn)?shù)偏微積分分類數(shù)學(xué)模型的關(guān)聯(lián)挖掘方法?;谄⒎e分原理塑造基于偏微積分方程的融合算法模型,實(shí)現(xiàn)大數(shù)據(jù)分類過(guò)程中的差異性數(shù)據(jù)融合;再通過(guò)偏微分分類數(shù)學(xué)模型的雙邊界收斂控制,在數(shù)據(jù)集合融入偏微積分分類數(shù)據(jù)模型,通過(guò)增減量支持向量完成數(shù)據(jù)的模糊控制,采用約束捆綁聚類算法對(duì)數(shù)據(jù)模型實(shí)施挖掘,獲取子序列,在最小迭代次數(shù)和收斂下,通過(guò)測(cè)度信息調(diào)控,采用高斯核函數(shù)挖掘關(guān)聯(lián)數(shù)據(jù)序列。實(shí)驗(yàn)結(jié)果說(shuō)明,所提關(guān)聯(lián)數(shù)據(jù)挖掘方法具有較高的挖掘效率和精度,穩(wěn)定性強(qiáng)。
關(guān)鍵詞: 偏微積分分類; 數(shù)學(xué)模型; 關(guān)聯(lián)挖掘; 分?jǐn)?shù)階; 收斂控制; 挖掘效率
中圖分類號(hào): TN911?34 文獻(xiàn)標(biāo)識(shí)碼: A 文章編號(hào): 1004?373X(2018)13?0095?05
Abstract: An association mining method based on fractional partial calculus classification mathematical model is put forward to improve the efficiency and accuracy of association data mining under the environment of big data mining. On the basis of partial calculus principle, the fusion algorithm model based on partial calculus equations is constructed to realize the difference data fusion in the large data classification process. By means of the dual?boundary convergence control of partial differential classification mathematical model, the data set is integrated into the data model of partial calculus classification. The variation of support vector is used to realize the fuzzy control of data. The constraint bundling clustering algorithm is used to mine the data model to obtain the sub sequences. Under the conditions of minimum iteration times and convergence, the Gaussian kernel function is used to mine the association data sequence by means of measuring information control. The experimental results show that the proposed association data mining method has high mining efficiency and accuracy, and strong stability.
Keywords: partial calculus classification; mathematical model; association mining; fractional order; convergence control; mining efficiency
當(dāng)前社會(huì)的信息化水平不斷提升,形成了海量的大數(shù)據(jù),大數(shù)據(jù)分類問(wèn)題成為不同領(lǐng)域研究的熱點(diǎn)問(wèn)題。高效的大數(shù)據(jù)關(guān)聯(lián)挖掘方法,為人們尋求有價(jià)值的信息提供基礎(chǔ),對(duì)于提升社會(huì)的信息化進(jìn)程具有重要應(yīng)用價(jià)值。隨著計(jì)算數(shù)學(xué)研究領(lǐng)域的不斷擴(kuò)張,分析偏微積分方程的穩(wěn)定解以及收斂性問(wèn)題逐漸引起人們的關(guān)注[1]。因此,本文提出基于分?jǐn)?shù)偏微積分分類數(shù)學(xué)模型的關(guān)聯(lián)挖掘方法,提高大數(shù)據(jù)環(huán)境下關(guān)聯(lián)數(shù)據(jù)挖掘的效率和精度。
1.1 基于偏微積分分類融合算法的數(shù)學(xué)模型
當(dāng)前在關(guān)聯(lián)數(shù)據(jù)挖掘領(lǐng)域中廣泛采用偏微積分原理,其能夠提高關(guān)聯(lián)數(shù)據(jù)的高頻區(qū)域,動(dòng)態(tài)存儲(chǔ)數(shù)據(jù)的低頻區(qū)域,使得數(shù)據(jù)的干擾因素增加。而偏微積分原理提升數(shù)據(jù)低頻區(qū)域時(shí),存儲(chǔ)數(shù)據(jù)的最低頻區(qū)域,其對(duì)階次的選擇要求較高[2]。如果采用小階次將降低干擾效果,采用大階次會(huì)形成模糊問(wèn)題。偏微積分原理解決離散數(shù)據(jù)過(guò)程中,無(wú)法處理待挖掘數(shù)據(jù)中噪聲的干擾問(wèn)題。本文在關(guān)聯(lián)數(shù)據(jù)挖掘過(guò)程中采用偏微積分原理,塑造關(guān)聯(lián)數(shù)據(jù)挖掘模型,實(shí)現(xiàn)基于偏微積分原理的差異性數(shù)據(jù)融合,提高關(guān)聯(lián)數(shù)據(jù)挖掘效率。
1.1.1 偏微積分方程
偏微積分方程是由整數(shù)階偏微分方程的轉(zhuǎn)化產(chǎn)生的,偏導(dǎo)數(shù)是將整數(shù)階微分方程中對(duì)函數(shù)影響因子的偏導(dǎo)數(shù)項(xiàng)進(jìn)行替換得到[3]。偏微積分方程為:
針對(duì)大數(shù)據(jù)環(huán)境下的關(guān)聯(lián)數(shù)據(jù)挖掘問(wèn)題,本文提出基于偏微積分分類數(shù)學(xué)模型的關(guān)聯(lián)數(shù)據(jù)挖掘方法。實(shí)驗(yàn)證明該方法提高了數(shù)據(jù)挖掘效率以及精度,獲得了令人滿意的效果。
[1] 潘大勝,陳志福,覃煥昌.基于模糊關(guān)聯(lián)迭代分區(qū)的挖掘優(yōu)化方法研究[J].科學(xué)技術(shù)與工程,2016,16(24):235?238.
PAN Dasheng, CHEN Zhifu, QIN Huanchang. Research on mining optimization based on fuzzy association iterative partition [J]. Science technology engineering, 2016, 16(24): 235?238.
[2] 馬瑞,周謝,彭舟,等.考慮氣溫因素的負(fù)荷特性統(tǒng)計(jì)指標(biāo)關(guān)聯(lián)特征數(shù)據(jù)挖掘[J].中國(guó)電機(jī)工程學(xué)報(bào),2015,35(1):43?51.
MA Rui, ZHOU Xie, PENG Zhou, et al. Considering the data mining of statistical parameters of temperature factors associa?ted feature data mining [J]. Proceedings of the CSEE, 2015, 35(1): 43?51.
[3] 周發(fā)超,王志堅(jiān),葉楓,等.關(guān)聯(lián)規(guī)則挖掘算法Apriori的研究改進(jìn)[J].計(jì)算機(jī)科學(xué)與探索,2015,9(9):1075?1083.
ZHOU Fachao, WANG Zhijian, YE Feng, et al. Research on association rules mining algorithm Apriori improvement [J]. Computer science and exploration, 2015, 9(9): 1075?1083.
[4] 張松,張琳.一種數(shù)據(jù)挖掘中的W?PAM限制聚類算法[J].計(jì)算機(jī)科學(xué),2016,43(z2):447?450.
ZHANG Song, ZHANG Lin. A W?PAM constrained clustering algorithm in data mining [J]. Computer science, 2016, 43(S2): 447?450.
[5] 徐開(kāi)勇,龔雪容,成茂才.基于改進(jìn)Apriori算法的審計(jì)日志關(guān)聯(lián)規(guī)則挖掘[J].計(jì)算機(jī)應(yīng)用,2016,36(7):1847?1851.
XU Kaiyong, GONG Xuerong, CHENG Maocai. The audit log association rules mining algorithm based on improved Apriori [J]. Computer applications, 2016, 36(7): 1847?1851.
[6] 劉自然,王律強(qiáng),李愛(ài)民,等.改進(jìn)Apriori算法對(duì)試車臺(tái)監(jiān)測(cè)數(shù)據(jù)的關(guān)聯(lián)挖掘[J].中國(guó)測(cè)試,2015,41(4):106?109.
LIU Ziran, WANG Lüqiang, LI Aimin, et al. Improve the association of the Apriori algorithm to the monitoring data of the test platform [J]. China test, 2015, 41(4): 106?109.
[7] 胡維華,馮偉.基于分解事務(wù)矩陣的關(guān)聯(lián)規(guī)則挖掘算法[J].計(jì)算機(jī)應(yīng)用,2014,34(z2):113?116.
HU Weihua, FENG Wei. Association rule mining algorithm based on decomposition transactional matrix [J]. Computer applications, 2014, 34(S2): 113?116.
[8] 李濤,林陳,王麗娜.一種改進(jìn)的相關(guān)項(xiàng)對(duì)挖掘算法研究[J].計(jì)算機(jī)仿真,2016,33(8):223?228.
LI Tao, LIN Chen, WANG Lina. An improved correlation for the research of mining algorithms [J]. Computer simulation, 2016, 33(8): 223?228.
[9] 黃立鋒,鄧玉輝.可時(shí)間局部性感知的塊I/O關(guān)聯(lián)挖掘算法[J].小型微型計(jì)算機(jī)系統(tǒng),2015,36(5):990?995.
HUANG Lifeng, DENG Yuhui. Temporal locally sexy block I/O association mining algorithm [J]. Minicomputer system, 2015, 36(5): 990?995.
[10] 王英博,馬菁,柴佳佳,等.基于Hadoop平臺(tái)的改進(jìn)關(guān)聯(lián)規(guī)則挖掘算法[J].計(jì)算機(jī)工程,2016,42(10):69?74.
WANG Yingbo, MA Jing, CHAI Jiajia, et al. Improved association rules mining algorithm based on Hadoop platform [J]. Computer engineering, 2016, 42(10): 69?74.