Search

Asocijativna pravila pogodna su za pronalaženje relacijskih odnosa između više varijabli u velikim skupovima podataka. Odnosi između varijabli, takozvana asocijativna pravila, izvode se u obliku implikacija AKO-ONDA. Najčešće korišteni algoritmi generiranja asocijativnih pravila su Apriori, FP-Growth, Tertius, Eclat i OPUS. U okviru ovog rada opisani su principi rada algoritama Apriori i FP-Growth te je opisan postupak implementacija algoritma Apriori unutar sustava za upravljanje podatcima imena Splunk. U konačnici vlastita implementacija algoritma Apriori uspoređena je na više skupova podataka sa postojećom implementacijom istog algoritma u aplikaciji imena Weka. The association rules are suitable for finding relationships between multiple variables in large sets of data. Relationships between variables, so-called association rules, are derived in the form of IF-THEN implications. The most commonly used algorithms for generating association rules are Apriori, FP- Growth, Tertius, Eclat and OPUS. This thesis describes the principles of Apriori and FP-Growth algorithms and also describes the procedure of Apriori algorithm implementation within Splunk data management system. Results of Splunk implementation of the Apriori algorithm are compared to multiple data sets with the existing implementation of the same algorithm in the Weka application.

One of the most popular methods of knowledge discovery in databases is the extraction of association rules. There are many different algorithms for association rule learning , which differ in space and time complexity. To perform a comparative analysis, we have implemented Apriori, Eclat and FP-growth algorithms and compared their time and memory consumption using synthetic and real databases. The analysis has shown that the FP-growth algorithm is the most efficient in the majority of cases. Zanimanje za metode odkrivanja znanja v podatkovnih bazah nenehno raste. Sodobne podatkovne baze so zelo velike, dosegajo terabajte in težijo k nadaljnemu povečanju, kar zahteva učinkovite, razširljive algoritme, ki lahko rešujejo težavo obdelave podatkov v razumnem roku. Ena od najbolj učinkovitih in priljubljenih metod za odkrivanje znanja je učenje asociacijskih pravil za ugotovitev različnih vrst pravilnosti v podatkih.

Association rule mining is one of the most important fields of data mining and knowledge discovery. In this paper, we propose an efficient multiple support frequent pattern growth algorithm which we called “MSFP-growth” that enhancing the FPgrowth algorithm by making infrequent child node pruning step with multiple minimum support using maximum constrains. The algorithm is implemented, and it is compared with other common algorithms: Apriori-multiple minimum supports using maximum constraints and FP-growth. The experimental results show that the rule mining from the proposed algorithm are interesting and our algorithm achieved better performance than other algorithms without scarifying the accuracy., {"references":["S. Brin, R. Motwani, J. Ullman, and S. Tsur, \"Dynamic itemset counting\nand implication rules for market basket data,\" in SIGMOD-97, 1997, pp.\n255-264.","C. Aggarwal, and P. Yu \"Online generation of association rule,\" in\nICDE-98, (1998), pp. 402-411.","R. Uday Kiran, P. Krishna Reddy , \"An efficient approach to mine rare\nassociation rules using maximum items support constraints data Security\nand security data, \" Springer, vol. 6121, pp. 84-95,2010.","L. Troiano, G. Scibelli and C. Birtolo, \"A fast algorithm for Mining rare\nitemsets,\" in Conf. 2009 the Ninth IEEE International Conference on\nIntelligent Systems Design and Applications, pp. 1149-1155.","K. S. C. Sadhasivam, T. Angamuthu, \"Mining rare Itemset with\nAutomated Support Thresholds,\" Computer Science J., vol. 7, pp. 394-\n399, 2011.","R. Agrawal, T. Imielinksi and A. Swami, \"Mining association rules\nbetween sets of items in large database,\" in 1993The ACM SIGMOD\nConf., pp. 207–216.","R. Agrawal and R. Srikant, \"Fast algorithms for mining association\nrules,\" in 1994VLDB Conf., Santiago, Chile, Expanded version\navailable as IBM Research Report RJ9839.","Y. Lee, T. Hong and W. Lin, \"Mining association rules with multiple\nminimum supports using maximum constraints,\" Elsevier Inc., 2005.","B. Liu, W. Hsu and Y. Ma, \"Mining association rules with multiple\nminimum supports,\" in 1999The International Conf. on Knowledge\nDiscovery and Data Mining, pp. 337–341.\n[10] J. Han, J. Pei and Y. Yin, \"Mining frequent patterns without candidate\ngeneration,\" in 2000International Conf. on Management of data, vol. 29,\npp. 1-12.\n[11] Z. Sun, \"Analysis and implementation of the algorithm of fp-growth,\"\nGuangxi Institute of Technology J., vol. 16, no. 3, pp. 64-67, 2004.\n[12] R. U. Kiran, P. K. Reddy, \"An improved multiple minimum support\nbased approach to mine rare association rules,\" in 2009 IEEE Int. Conf.\nCommunications Computational Intelligence and Data Mining, pp. 340-\n347.\n[13] R. U. Kiran, P. K. Reddy , \"An Improved Frequent Pattern growth\nApproach To Discover Rare Association rules,\" in 2009 International\nConf. on Knowledge Discovery and Information Retrieval, pp. 43-52.\n[14] S. Tsang, Y. S. Koh and G. Dobbie, \"Finding Interesting Rare\nAssociation Rules Using Rare Pattern Tree,\" The University of\nAuckland, Springer-Verlag Berlin Heidelberg,2013.\n[15] J. Pei, J. Han, B. Mortazavi-Asl, J. Wang, H. Pinto, Q. Chen, M. Hsu, ,\n\"Mining sequential patterns by pattern-growth: the prefixspan\napproach,\" IEEE Trans. Knowledge and Data Engineering 16, 1424–\n1440, 2004.\n[16] R. Srikant, Q. Vu, and R. Agrawal, \"Mining association rules with item\nconstraints,\" KDD-97, pp. 67-73, 1997.\n[17] Y. Hu, F. Wu, Y. Liao, \"Sequential pattern mining with multiple minimum\nsupports: a tree based approach\", in 2010, The 2nd International\nConf. on Software Engineering and Data Mining, Chengdu, China.\n[18] J. Pie, J. Han, and L. Lakshmanan., \"Mining frequent itemsets with\nconvertible constraints\", in 2001 IEEE ICDE Conf., p.433– 442.\n[19] M. S. Saravanan, R.J. Rama, \"Performance study of Association Rule\nMining Algorithms for Dyeing Processing System Innovative Systems\nDesign and Engineering\", Vol 2, No 5, 2011.\n[20] http://www.cs.waikato.ac.nz/ml/weka , Last acess : 6 march 2015 \n[21] M. Walaa, H. Ahmed and K. Hoda, \"Combined Algorithm for Data\nMining using Association rules,\" Ain Shams of Electrical Engineering\nJ., Vol. 1.No. 1 ISSN: 1687-8582, 2008.\n[22] P. Jyothi, V. D. Mytri, \"A Fast association rule algorithm based on\nbitmap computing with multiple minimum supports using max\nconstraints,\" International of Comp. Sci & Electronics J.,Volume1,Issue\n2, IJCSEE,2013.\n[23] F.A. Hoque , N. Easmin and K. Rashed, \" Frequent pattern mining for\nmultiple minimum supports with support tuning and tree maintenance on\nincremental database,\" Research of Information Technology J., 3(2): 79-\n90 ,2012."]}

This article proposes an innovative utility sentient approach for the mining of interesting association patterns from transaction databases. First, frequent patterns are discovered from the transaction database using the FP-Growth algorithm. From the frequent patterns mined, this approach extracts novel interesting association patterns with emphasis on significance, utility, and the subjective interests of the users. The experimental results portray the efficiency of this approach in mining utility-oriented and interesting association rules. A comparative analysis is also presented to illustrate our approach's effectiveness.

Actualmente, a prescrição electrónica na prática clínica é muitas vezes utilizada em acções de apoio à tomada de decisão e à monitorização dos serviços prestados por organismos ligados à saúde pública. O sucessivo armazenamento de dados relativos à prescrição de medicamentos fez com que, já hoje, se possa aplicar algumas técnicas de mineração de dados com sucesso na procura de padrões de prescrição de medicamentos que possam ser interessantes para a optimização dos serviços dos organismos em causa. As regras de associação, usualmente utilizadas em processos de análise sobre dados referentes a transacções, podem ser aplicadas nesta área com alguma pertinência e utilidade. Na realidade, a ideia de descobrir padrões de prescrição de medicamentos é, só por si, muito atractiva pelo leque de novas opções de gestão e de rentabilização de serviços que abre. Neste trabalho apresentamos os resultados de um estudo realizado sobre a prescrição de medicamentos na região norte de Portugal, cujo objectivo principal foi o de descobrir padrões de prescrições entre os medicamentos prescritos, os médicos prescritores e os laboratórios prescritos., Today, electronic prescription in clinical scenarios is frequently used to support decision-making tasks and services monitoring in public health institutions. The huge amount of data that has been stored during last years about medication prescriptions makes now possible to apply data mining techniques with a significant success to discover patterns that could be useful to optimize medical services and resources related to medication prescription. Association rules, which are frequently used in the analysis of commercial transactions data, could be applied usefully on the medical prescriptions area. In fact, the idea of discovering medication prescriptions patterns it is quite attractive due to the wide range of new applications that can be opened to service management and optimisation. In this work, we present a study about medication prescriptions in the north region of Portugal, in order to discover some relevant patterns among medication, doctors and laboratories., (undefined)

Journal of information systems in Business JISB, 6, 3, 56