Summary of Computer doctoral thesis: Mining fuzzy association rules and fuzzy sequential patterns in temporal quantitative databases

The objective of the thesis: Mining association rules with time-interval between events in temporal quantitative databases called fuzzy time-interval association rules; mining sequential patterns with time-interval between events in temporal quantitative sequential databases called fuzzy sequential patterns with fuzzy time intervals; mining common sequential rules with time-interval between events in temporal quantitative sequential databases called fuzzy common sequential rules with fuzzy time intervals.
Nội dung trích xuất từ tài liệu:
Summary of Computer doctoral thesis: Mining fuzzy association rules and fuzzy sequential patterns in temporal quantitative databases MINISTRY OF EDUCATION AND VIETNAM ACADEMY TRAINING OF SCIENCE AND TECHNOLOGY GRADUATE UNIVERSITY SCIENCE AND TECHNOLOGY ---------------------------- TRUONG DUC PHUONG MINING FUZZY ASSOCIATION RULES AND FUZZY SEQUENTIAL PATTERNS IN TEMPORAL QUANTITATIVE DATABASES Major: Information Systems Code: 9 48 01 04 SUMMARY OF COMPUTER DOCTORAL THESIS Ha Noi – 2021 The thesis has been completed at Graduate University Of Science And Technology - Vietnam Academy Of Science And Technology Supervisor 1. Assoc. Prof. Dr. Do Van Thanh Supervisor 2. Assoc. Prof. Dr. Nguyen Duc Dung Review 1: Review 2: Review 3: The thesis will be defended at the Board of Examiners of Graduate University Of Science And Technology - Vietnam Academy Of Science And Technology, at……………….on…………………………. The thesis can be explored at: - Library of Graduate University Of Science And Technology - National Library of Vietnam INTRODUCTION 1. Motivation of the thesis (Phương and Thành, 2013) Mining association rules and sequential patterns, sequential rules are some of the most important domains in data mining. Up to now, a lot of research related to them. Association rules and sequential patterns, sequential rules are proposed in many forms such as transaction/quantitative; weighted / unweighted; with / without time; etc. Rekesh Agrawal et al first introduced an association rule mining problem in transaction databases in 1993 (Agrawal, Imieliński and Swami, 1993) and up to now, there have been many proposed algorithms according to many different approaches to mining the rules in transaction databases such as APRIORI (Agrawal, Srikant and others, 1994), PARTITION (Savasere, Omiecinski and Navathe, 1995), A-CLOSE (Pasquier et al., 1999a), A-CLOSE+ (Shekofteh, Rahmani and Dezfuli, 2008), CLOSE (Pasquier et al., 1999b), CLOSET (Pei et al., 2000), CLOSET+ (Wang, Han and Pei, 2003), CHARM (Zaki and Hsiao, 2002), MAFIA (Burdick, Calimlim and Gehrke, 2001), GENMAX (Gouda and Zaki, 2005), ECLAT (Ogihara et al., 1997), DIC (Brin et al., 1997), FP-GROWTH (Han et al., 2004), CFPMINE (Qin, Luo and Shi, 2004), ETARM (Nguyen et al., 2018), LRM (Saravanan and Sree, 2011), PARM (Sumathi and Kirubakaran, 2012), NEGFIN (Aryabarzan, Minaei-Bidgoli and Teshnehlab, 2018). However, there are many databases in which values of attributes are numeric or categorical called quantitative databases. Mining association rules in a quantitative database often uses one of two ways: discretized (Srikant and Agrawal, 1996a; Lent, Swami and Widom, 1997; Fukuda et al., 1999; Rastogi and Shim, 2002) and fuzzy values of quantitative attributes (Chan and Au, 1997; Kuok, Fu and Wong, 1998; T.-P. Hong, Kuo and Chi, 1999; Hong, Kuo and Chi, 2001; Hong, Chiang and Wang, 2002; Hong, 2003). The essence of the first approach is transform quantitative database into the transaction database by converting the qualitative attributes into a number of corresponding items and then applying one of the mining association rules algorithms in transaction databases. The second approach solves the disadvantages of the first, but they require need to be improved the algorithms. Temporal database is a database that stores information about the time of transactions (Tansel et al., 1993) (Aydin and Angryk, 2018). In 1998, Lu et al. (Lu, Han and Feng, 1998) proposed association rule with time interval between transactions in the temporal databases. The rules had a form → where a, b are itemsets. In work (Lu, Han and Feng, 1998), two algorithms, E-Apriori and EH-Apriori, were proposed. On the main idea, the two algorithms E-Apriori, EH-Apriori are based on the idea of the Apriori algorithm and use a sliding window for the time interval. For mining association rules with time interval, further algorithms are proposed such as FITI (Tung et al., 2003), ITARM (Qin and Shi, 2006), ITP-Miner (Lee and Wang, 2007), IAR Miner (Nandagopal, Arunachalam and Karthik, 2012), CITP-Miner (Nguyen et al., 2019), NCITPS-MINER (Nguyen et al., 2020). Mining association rule with time interval has only applied for temporal transaction databases, but has not yet been done for temporal quantitative databases. This is a research gap that the thesis wishes to solve. Sequential rule, sequential pattern are known as the classical sequential rule, classical sequential pattern to distinguish them from a kind of new sequential rule, sequential pattern proposed in recent years. The classical sequential patterns (referred to as the sequential pattern) are classical sequences (in short, sequences) that are frequent in sequential databases. The sequential patterns show relationships between transactions in the sequence. Mining sequential pattern in sequential databases was the first introduced in 1995 (Agrawal, Srikant and others, 1995) and continues to receive a lot of attention. There are many algorithms to mine sequential patterns in sequential transaction databases such as GSP (Srikant and Agrawal, 1996b), SPIRIT (Garofalakis, Rastogi and Shim, 1999), SPADE (Zaki, 2001), 1 SPAM (Ayres et al., 2002), FAST (Salvemini et al., 2011), CM-SPADE (Fournier-Viger, Gomariz, Campos, et al., 2014), MAXSP (Fournier-Viger, Wu and Tseng, 2013), GENMINER (Lo, Khoo and Li, 2008), FREESPAN (Han et al., 2000), PREFIXSPAN (Pei et al., 2004), CLOSPAN (Yan, Han and Afshar, 2003), MSPIC-DBV (Van, Vo and Le, 2018), HSPREC (Bhatta, Ezeife and Butt, 2019),... Temporal sequential database is a sequential database which include infomation ...