周亮
Abstract: In order to improve the ability of big data mining in weapon information management database, a big data mining technique based on semantic similarity association feature extraction is proposed. The distributed structure model of weapon equipment information management data is constructed, the association rule information fusion method is used to deal with the big data block structure matching, and the adaptive regression analysis method is used to extract the association feature of the weapon equipment information management data. The association rule features of the equipment management data are classified and identified, the semantic similarity of big data is calculated, and the big data fusion scheduling is carried out by combining the decision statistical analysis method. The semantic partition method is used to deal with big data fuzzy clustering to realize the big data adaptive mining performance in the weapon equipment information management database. The simulation results show that this method has high accuracy and strong anti-interference ability to redundant data, the adaptive mining and retrieval ability of weapon equipment big data is improved.
引言
在信息化作戰(zhàn)條件下,海軍武器裝備信息數(shù)據(jù)庫的規(guī)模不斷增大,在武器裝備信息數(shù)據(jù)庫中融合了大量的裝備管理信息數(shù)據(jù),常見的武器裝備信息數(shù)據(jù)主要有武器性能數(shù)據(jù)、武器作戰(zhàn)狀態(tài)信息數(shù)據(jù)、作戰(zhàn)人員信息數(shù)據(jù)以及維修管理信息數(shù)據(jù),大量的武器裝備信息數(shù)據(jù)融合在數(shù)據(jù)庫中,構(gòu)建了海軍武器裝備大數(shù)據(jù)庫。為了提高信息化作戰(zhàn)能力,需要對武器裝備信息大數(shù)據(jù)進(jìn)行優(yōu)化挖掘,結(jié)合數(shù)據(jù)庫檢索和自適應(yīng)信息融合方法,提高武器裝備信息數(shù)據(jù)的自適應(yīng)調(diào)度和檢索性能,研究信息化作戰(zhàn)條件下的武器裝備信息大數(shù)據(jù)挖掘算法在提高裝備信息管理方面具有重要意義[1]。
當(dāng)前,對武器裝備信息庫中的大數(shù)據(jù)的挖掘方法主要有關(guān)聯(lián)規(guī)則挖掘方法、閉頻繁項(xiàng)集挖掘方法、模糊C均值挖掘方法和時頻特征提取挖掘方法等[2-3],構(gòu)造武器裝備信息庫中的大數(shù)據(jù)特征分布的模糊聚類中心,采用自適應(yīng)的數(shù)據(jù)分類和信息融合方法進(jìn)行大數(shù)據(jù)挖掘,取得了較好的數(shù)據(jù)挖掘效果。其中,文獻(xiàn)[4]中提出一種基于模糊C均值聚類的武器裝備信息庫中的大數(shù)據(jù)挖掘方法,構(gòu)建大數(shù)據(jù)的分類調(diào)度模型,采用屬性關(guān)聯(lián)規(guī)則特征提取方法進(jìn)行大數(shù)據(jù)關(guān)聯(lián)特征提取,實(shí)現(xiàn)裝備信息管理數(shù)據(jù)的優(yōu)化挖掘,提高數(shù)據(jù)挖掘的準(zhǔn)確性和自適應(yīng)性,但該方法進(jìn)行大數(shù)據(jù)挖掘的計(jì)算開銷較大,自適應(yīng)性能不好。文獻(xiàn)[5]中提出一種基于模糊分區(qū)聚類的武器裝備信息庫中的大數(shù)據(jù)關(guān)聯(lián)挖掘改進(jìn)算法,構(gòu)建判別統(tǒng)計(jì)量進(jìn)行大數(shù)據(jù)的自適應(yīng)融合處理,實(shí)現(xiàn)對武器裝備信息數(shù)據(jù)的優(yōu)化挖掘,但該方法進(jìn)行大數(shù)據(jù)挖掘的自適應(yīng)分類調(diào)度性能不好,抗干擾能力不強(qiáng)。
針對上述問題,本文提出一種基于語義相似性關(guān)聯(lián)特征提取的大數(shù)據(jù)挖掘技術(shù)。首先構(gòu)建武器裝備信息管理數(shù)據(jù)分布式結(jié)構(gòu)模型,采用關(guān)聯(lián)規(guī)則信息融合方法進(jìn)行大數(shù)據(jù)分塊結(jié)構(gòu)匹配處理,結(jié)合自適應(yīng)回歸分析方法進(jìn)行武器裝備信息管理數(shù)據(jù)的關(guān)聯(lián)特征提取,對提取的裝備管理數(shù)據(jù)的關(guān)聯(lián)規(guī)則特征量進(jìn)行屬性分類識別。然后計(jì)算大數(shù)據(jù)的語義相似性關(guān)聯(lián)特征量,結(jié)合判決統(tǒng)計(jì)分析方法進(jìn)行大數(shù)據(jù)的融合調(diào)度,采用語義劃分方法進(jìn)行大數(shù)據(jù)模糊聚類處理,實(shí)現(xiàn)武器裝備信息管理數(shù)據(jù)庫中的大數(shù)據(jù)自適應(yīng)挖掘。最后通過仿真實(shí)驗(yàn)進(jìn)行性能測試,展示了本文方法在提高大數(shù)據(jù)準(zhǔn)確挖掘能力方面的優(yōu)越性能。