随着我国航天事业的不断发展，航天器平台有效载荷的作用越来越重要。研究有效载荷之间的关联关系，对卫星运行性能分析、异常排除、故障诊断、反馈卫星设计具有重要意义。论文针对现有有效载荷数据分析算法效率低、相关性挖掘不充分的问题，进行了以下研究：

(1)针对相关性分析算法对高维、模糊的有效载荷数据关联关系挖掘不充分的问题，提出了基于模糊灰色关联系数的载荷数据相关性分析算法。首先考虑数据之间的模糊性，基于斯皮尔曼相关系数改进原有隶属度为模糊隶属度，在此基础上构建模糊灰色相关性模型；其次根据模糊灰色关联度得到属性之间的相关序列，相较于原始序列，更加清晰、层次分明地展示有效载荷各属性之间的相关性程度；最后与传统相关性系数进行了对比实验。实验结果表明模糊灰色关联度相较于肯德尔相关系数、斯皮尔曼相关系数、皮尔逊相关系数和最大信息系数，平均关联程度提高了0.623、0.425、0.621、0.115，由此验证了论文提出的相关性分析算法能够进一步分析数据之间的模糊性和相关性，提升了数据的可用性，为后续关联规则的提取奠定了基础。

(2)针对关联规则算法运行效率低、占用内存大的问题，提出了基于二次随机抽样和哈希桶的卫星载荷关联规则算法(RS_Hash)。首先对卫星载荷数据进行预处理，之后使用二次随机抽样的方法获取最佳样本数据，并设计抽样误差和抽样停止规则确定最合适的样本容量；其次基于哈希函数映射数据的思想，使用哈希桶存储样本中的频繁项集，并通过挖掘的频繁项集生成关联规则；最后在卫星载荷数据集上进行实验，并和Apriori、PCY、SON、FP-Growth、RCM_Apriori和Hash_Cumulate算法进行对比。实验结果表明，在不同事务长度和支持度的情况下，相较于其它六种相关算法，RS_Hash算法在平均时间效率上分别提高了75.81%、49.10%、59.38%、50.22%、40.16%和39.22%，同时减少了内存的占用。

(3)针对从卫星载荷数据提取的关联规则存在冗余的问题，提出了基于规则更新的深度嵌入聚类算法(CADC)。首先，将数据之间的模糊灰色关联系数添加到自编码器网络的嵌入层，将其放入自编码器的预训练中，在数据降维的同时尽可能的保存原始输入的相关性信息；其次，在聚类空间中使用模糊灰色关联系数确定初始聚类中心，并在聚类时提出了规则更新算法进行迭代，通过网络训练获得一个理想的聚类结果；最后在不同的数据集上对CADC算法进行了实验。实验结果表明CADC算法相较于DEC、IDCE、DCN、VaDE、DCEC、DEPICT算法，在卫星载荷数据集上的聚类轮廓系数分别提高了7.8%、7.4%、14.6%、6.9%、10.2%和9.6%，取得了更好的聚类效果，有效地减少了冗余的关联规则。

With the continuous development of Chinese space industry, the function of payload of spacecraft platform is more and more important. It is of great significance to study the correlation between payloads for satellite performance analysis, anomaly removal, fault diagnosis and feedback satellite design. Aiming at the problems of low efficiency and insufficient correlation mining of existing payload data analysis algorithms, the following researches are carried out:

(1) Aiming at the problem that the traditional correlation analysis algorithm is not enough to mine the correlation relation of high-dimensional and fuzzy payload data, a new correlation analysis algorithm based on fuzzy grey correlation coefficient is proposed. Firstly, considering the fuzziness between the data, the original membership degree is improved to fuzzy membership degree based on Spearman correlation coefficient, and the fuzzy grey correlation model is constructed on this basis. Secondly, the correlation sequence between attributes is obtained according to the fuzzy gray correlation degree. Compared with the original sequence, the correlation degree between payload attributes is more clearly and clearly displayed. Finally, the comparison experiment with traditional correlation coefficient is carried out. The experimental results show that compared with Kendall correlation coefficient, Spearman correlation coefficient, Pearson correlation coefficient and maximum information coefficient, the average correlation degree of fuzzy grey correlation is increased by 0.623, 0.425, 0.621 and 0.115, which verifies that the correlation analysis algorithm proposed in this paper can further analyze the fuzziness and correlation between data. It improves the availability of data and lays a foundation for the subsequent extraction of association rules.

(2) Aiming at the problems of low efficiency and large memory consumption of traditional association rule algorithm, a satellite load association rule algorithm (RS_Hash) based on quadratic random sampling and hash bucket was proposed. Firstly, the satellite load data was preprocessed, and then the optimal sample data was obtained by quadratic random sampling, and the sampling error and sampling stop rules were designed to determine the most appropriate sample size. Secondly, based on the idea of hash function mapping data, the hash bucket is used to store frequent item sets in the sample, and the association rules are generated through the frequent item sets mined. Finally, the Apriori, PCY, SON, FP-Growth, RCM_Apriori and Hash_Cumulate algorithms were compared on the satellite payload data set. The experimental results show that compared with the other six related algorithms, RS_Hash algorithm improves the average time efficiency by 75.81%, 49.10%, 59.38%, 50.22%, 40.16% and 39.22%, respectively, and reduces the memory consumption under different transaction length and support degree.

(3) Aiming at the problem of redundancy in association rules of satellite payload data, a deep embedded clustering algorithm (CADC) based on rule update is proposed. First, add the fuzzy gray correlation coefficient between the data to the embedding layer of the autoencoder network, put it into the pre-training of the autoencoder, and preserve the correlation information of the original input as much as possible while reducing the data dimension; Secondly, use the fuzzy gray correlation coefficient to determine the initial cluster center in the cluster space, and propose a rule update algorithm to iterate during clustering, and obtain an ideal clustering result through network training; finally, on different data sets The CADC algorithm was tested. The experimental results show that compared with the DEC, IDCE, DCN, VaDE, DCEC, and DEPICT algorithms, the clustering silhouette coefficients of the CADC algorithm on the satellite payload data set are increased by 7.8%, 7.4%, 14.6%, 6.9%, 10.2% and 9.6%, achieved a better clustering effect，the redundant association rules are effectively reduced.

[1]程鹏杰. 基于遥测数据的卫星动量轮故障检测与预测研究[D].北京交通大学,2022.

[2]蔡晓玮,智佳,陈志敏等.基于关联知识的航天器有效载荷遥测数据仿真方法[J].计算机工程与设计,2022,43(7):2095-2101.

[3]任国恒. 同步卫星遥测数据相关性分析与研究[D].西安工业大学,2011.

[4]崔阳,陈雪笛,姚小松等.遥感微纳卫星载荷平台高精度标定技术研究[J].南京航空航天大学学报,2022,54(3):499-507.

[5]孙宇豪,李国通,张鸽.一种基于相关概率模型的卫星异常检测方法[J].中国科学院大学学报,2021,38(3):409-416.

[6]孙宇豪. 基于多变量相关性分析的卫星异常检测技术研究[D].中国科学院大学(中国科学院微小卫星创新研究院),2020.

[7]A.F. Hayes, A. K. Montoya. A tutorial on testing, visualizing, and probing an interaction involving a multicategorical variable in linear regression analysis[J]. Communication Methods and Measures, 2017, 11(1): 1-30.

[8]严良涛.基于核的多元统计回归方法研究[D]. 南昌大学,2016.

[9]Mahmoudi M R, Heydari M H, Qasem S N, et al. Principal component analysis to study the relations between the spread rates of COVID-19 in high risks countries[J]. Alexandria Engineering Journal, 2021, 60(1): 457-464.

[10]Nobre J, Neves R F. Combining principal component analysis, discrete wavelet transform and XGBoost to trade in the financial markets[J]. Expert Systems with Applications, 2019, 125(5): 181-194.

[11]Dai M, Jin H, Kou S, et al. A dynamic mean-variance analysis for log returns[J]. Management Science, 2021, 67(2): 1093-1108.

[12]刘璐,王筱莉.基于改进熵权-灰色关联-TOPSIS的网络辟谣平台影响力研究[J].智能计算机与应用,2022,12(11):240-246.

[13]包顺,徐鑫,肖箭等.Pythagorean犹豫模糊灰色关联前景多属性决策方法[J].计算机工程与应用,2020,56(18):119-123.

[14]Yao R, Liu T, Huang X, et al. Multi-objective design of laminated cooling configuration applied in gas turbine based on fuzzy grey relational analysis[J]. International Journal of Heat and Mass Transfer, 2023, 207(25): 124-129.

[15]Khuman A S. The similarities and divergences between grey and fuzzy theory[J]. Expert Systems with Applications, 2021, 186(9): 115-121.

[16]Liu X, Liu H, Zhao X, et al. A novel neural network and grey correlation analysis method for computation of the heat transfer limit of a loop heat pipe (LHP)[J]. Energy, 2022, 259(17): 124-138.

[17]Dong Y, Qin S J. A nover dynamic PCA algorithm for dynamic data modeling and process monitoring[J]. Journal of Process Control.2018,67(14):1-11.

[18]Dong Y, Qin S J. Regression on dynamic PLS structures for supervised learning of dynamic data[J]. Journal of Process Control, 2018, 68(6): 64-72.

[19]Zhang M Q, Luo X L. Modified canonical variate analysis based on dynamic kernel decomposition for dynamic nonlinear process quality monitoring[J]. ISA transactions, 2021, 108(3): 106-120.

[20]Agrawal R, Srikant R. Fast algorithms for mining association rules in large database[C]. Proceeding 20th International Conference on Very Large Database. Santiago: Morgan Kaufmann Publishers Inc, 1994,16(5):487-499.

[21]Savasere A, Omiecinski E, Navathe S. An efficient algorithm for mining association rules in large databases[C].International Conference on Very Large Data Bases. Morgan Kaufmann Publishers Inc.2015,19(6):432-444.

[22]Nguyen L, Vo B, et al. ETARM: an efficient top-k association rule mining algorithm[J]. Applied Intelligence, 2018, 48(18): 1148-1160.

[23]Bhandari A, Gupta A, Das D . Improvised Apriori algorithm using frequent pattern tree for real time applications in data mining[J]. Procedia Computer Science, 2015, 46(8):644-651.

[24]毕玉萍,胡世昌,李劲华.基于排序树的Node-Apriori改进算法[J].青岛大学学报(自然科学版),2020,33(3):50-56.

[25]胡世昌.基于二进制编码的Apriori改进算法[J].计算机应用研究,2020,37(2):398-423.

[26]王伟,储泽楠,韩毅.基于MapReduce的Apriori前后项约束关联规则改进算法[J].信阳师范学院学报(自然科学版),2020,33(3):448-453.

[27]程江洲,闫冉阳,冯梦婷,冯馨以.基于ACT-Apriori算法的电网故障诊断方法研究[J].电子测量技术,2021,44(24):32-39.

[28]Han J, Pei J, Yin Y. Mining frequent patterns without candidate generation[J]. ACM SIGMOD Record, 2000, 29(2):1-12.

[29]陈志飞,冯钧. 一种基于Apriori算法的优化挖掘算法[J]. 计算机与现代化,2016(9):1-5.

[30]Li L,Li Q, Wu Y, et al. Mining association rules based on deep pruning strategies[J]. Wireless Personal Communications, 2017(2):1-25.

[31]于守健,周羿阳.基于前缀项集的Apriori算法改进[J].计算机应用与软件,2017,34(02):290-294.

[32]赵龙,杨小兵,吴强,等. 一种基于多值属性的改进Apriori算法[J]. 中国计量大学学报,2017,(1):108-112.

[33]吉祥,黄树成. 基于哈希树的并行关联规则挖掘算法研究[J].计算机与数字工程.2020,48(7):1601-1605.

[34]刘彦戎,杨云. 一种矩阵和排序索引关联规则数据挖掘算法[J].计算机技术与发展.2021,31(2):54-59.

[35]Shawkat M, Badawi M, El-ghamrawy S, et al. An optimized FP-growth algorithm for discovery of association rules[J]. The Journal of Supercomputing, 2022, 78(4): 5479-5506.

[36]Jang H J, Yang Y, Park J S, et al. FP-Growth algorithm for discovering region-based association rule in the IoT environment[J]. Electronics, 2021, 10(24): 3091.

[37]Jawthari M, Stoffa V. Relation between student engagement and demographic characteristics in distance learning using association rules[J]. Electronics, 2022, 11(5): 724.

[38]Liu H, Liu J. Female employment data analysis based on decision tree algorithm and association rule analysis method[J]. Scientific Programming, 2022, 25(3):1-11.

[39]Carmona-Pírez J, Poblador-Plou B, Poncel-Falcó A, et al. Applying the FAIR4Health solution to identify multimorbidity patterns and their association with mortality through a frequent pattern growth association algorithm[J]. International Journal of Environmental Research and Public Health, 2022, 19(4): 2040.

[40]Xue R, Tang P, Fang S. Prediction of computer network security situation based on association rules mining[J]. Wireless Communications and Mobile Computing, 2022,3(5) 1-9.

[41]张梓靖,张建勋,全文君等.多元时序的深度自编码器聚类算法[J/OL].计算机应用研究,2023,40(8):1-8.

[42]陶文彬,钱育蓉,张伊扬等.基于自编码器的深度聚类算法综述[J].计算机工程与应用,2022,58(18):16-25.

[43]林玮. 基于深度自编码器的属性网络异常检测算法研究[D].华南理工大学,2021.

[44]Xie J, Girshick R, FarhadA. Unsupervised deep embedding for clustering analysis[C]// International Conference on Machine Learning, 2016,50(7): 478-487.

[45]Guo X, Gao L, Liu X, etal. Improved deep embedded clustering with local structure preservation[C]//International Joint Conferences on Artificial Intelligence Organization, 2017,23(18): 1753-1759.

[46]Guo X, Liu X, Zhu E, et al. Deep clustering with convolutional autoencoders[C]// International Conference on Neural Information Processing, 2017,26(3): 373-382.

[47]Jiang Z, Zheng Y, Tan H, et al. Variational deep embedding: an unsupervised and generative approach to clustering[C]//Proceedings of the 26th International Joint Conference on Artificial Intelligence. 2017,32(16): 1965-1972.

[48]Cai J, Wang S, Guo W. Unsupervised embedded feature learning for deep clustering with stacked sparse autoencoder[J]. Expert Systems with Applications, 2021, 186(5): 115-123.

[49]Guo X, Liu X, Zhu E, et al. Deep clustering with convolutional autoencoders[C]// International Conference on Neural Information Processing，2017,24(26):373-382.

[50]Zheng Y, Jia C, Yu J, et al. Deep embedded clustering with distribution consistency preservation for attributed networks[J]. Pattern Recognition, 2023, 139(46): 109-118.

[51]Hafidi H, Ghogho M, Ciblat P, et al. Negative sampling strategies for contrastive self-supervised learning of graph representations[J]. Signal Processing, 2022, 190: 108310.

[52]Cao W, Zhang Z, Liu C, et al. Unsupervised discriminative feature learning via finding a clustering-friendly embedding space[J]. Pattern Recognition, 2022, 129: 108768.

[53]Wang C, Pan S, Hu R, et al. Attributed graph clustering: a deep attentional embedding approach[C]//Proceedings of the 28th International Joint Conference on Artificial Intelligence. 2019: 3670-3676.

[54]Dizaji K G，HerandiI A，Huang H. Deep clustering via joint convolutional autoencoder embedding and relative entropy minimization[C]//International Conference on Neural Information Processing, 2017: 5736-5745.

[55]Cai J，Wang S，Guo W. Unsupervised embedded feature learning for deep clustering with stacked sparse autoencoder[J]. Expert Systems with Applications, 2021,186: 115729.

[56]Wang C, Pan S, Celina P Y, et al. Deep neighbor-aware embedding for node clustering in attributed graphs[J]. Pattern Recognition, 2022, 122: 108230.

[57]Sun H, He F, Huang J, et al. Network embedding for community detection in attributed networks[J]. ACM Transactions on Knowledge Discovery from Data (TKDD), 2020, 14(3): 1-25.

[58]Zhang H, Li P, Zhang R, et al. Embedding graph auto-encoder for graph clustering[J]. IEEE Transactions on Neural Networks and Learning Systems, 2022: 1–11.

[59]Bo D, Wang X, Shi C, et al. Structural deep clustering network[C]// Proceedings of the web conference 2020. 2020: 1400-1410.

[60]Tu W, Zhou S, Liu X, et al. Deep fusion clustering network[C]// Proceedings of the AAAI Conference on Artificial Intelligence. 2021, 35(11): 9978-9987.

[61]Guo L, Dai Q. Graph clustering via variational graph embedding[J]. Pattern Recognition, 2022, 122: 108334.

[62]江刘锋. 相关系数在因果发现中的研究与应用[D].合肥工业大学,2022.

[63]赵源上,林伟芳.基于皮尔逊相关系数融合密度峰值和熵权法的典型新能源出力场景研究[J/OL].中国电力:1-10[2023-04-22].

[64]张琪. 基于肯德尔相关系数的函数型数据相关性检验[D].东北师范大学,2020.

[65]高明裕,蔡林辉,孙长城等.一种基于斯皮尔曼秩相关结合神经网络的电池组内部短路故障检测算法[J].电子与信息学报,2022,44(11):3734-3747.

[66]熊玲珠,邱伟涵,罗计根等.基于最大信息系数和迭代式XGBoost的混合特征选择方法[J].计算机应用与软件,2023,40(01):280-286+305.

[67]高山,智永锋,张普等.基于改进灰色关联分析的航天产品性能样机仿真结果一致性验证方法[J/OL].系统工程与电子技术,2023,45(9):1-9.

[68]董安国,张倩,刘洪超等.基于TSNE和多尺度稀疏自编码的高光谱图像分类[J].计算机工程与应用,2019,55(21):177-182+219.

[69]Miao Y, Wang J, Zhang B. Practical framework of Gini index in the application of machinery fault feature extraction. Mechanical Systems and Signal Processing,2022,165, 108333. 

[70]Shakeri M, Hesari M M. Discovering adjacency relationships of goods by using association rules to optimizing hypermarket layout[J]. Revista Turismo Estudos e Práticas-RTEP/UERN, 2020 (1): 1-12.

[71]He X, Teng S. Research on insurance marketing application based on hash link-table improved association rule algorithm[C]//2022 14th International Conference on Machine Learning and Computing (ICMLC). 2022: 126-133.

[72]Savasere, et al. Efficient association rule mining based SON algorithm for a big data platform[J]. Journal of Digital Contents Society, 2017, 18(8): 1593-1601.

[73]Han J, Pei J, Yin Y, et al. Mining frequent patterns without candidate generation: A frequent-pattern tree approach[J]. Data Mining & Knowledge Discovery, 2004, 8(1): 53-87.

[74]廖纪勇,吴晟,刘爱莲.基于布尔矩阵约简的Apriori算法改进研究[J].计算机工程与科学,2019,41(12):2231-2238.

[75]郭倩,殷丽凤.基于散列技术的多层关联规则算法的改进[J].计算机工程与设计,2021,42(9):2485-2491.

[76]Smith J S, Nebgen B T, Zubatyuk R, et al. Approaching coupled cluster accuracy with a general-purpose neural network potential through transfer learning[J]. Nature communications, 2019, 10(1): 2903.

[77]Rahmanian M, Mansoori E G. An unsupervised gene selection method based on multivariate normalized mutual information of genes[J]. Chemometrics and Intelligent Laboratory Systems, 2022, 222: 104512.

[78]Robert V, Vasseur Y, Brault V. Comparing high-dimensional partitions with the co-clustering adjusted rand index[J]. Journal of Classification, 2021, 38: 158-186.

[79]Dinh D T, Huynh V N. Estimating the optimal number of clusters in categorical data clustering by silhouette coefficient[C]//Knowledge and Systems Sciences: 20th International Symposium, 2019: 1-17.