International Journal of Engineering Research and Development e-ISSN: 2278-067X, p-ISSN: 2278-800X, www.ijerd.com Volume 11, Issue 07 (July 2015), PP.84-95 84 A Fuzzy Soft Set Approach for Mining Association Patterns Saakshi Saraf 1 , Neeru Adlakha 2 , Sanjay Sharma 3 1 Department Of Computer Applications,Maulana Azad National Institute Of Technology,Bhopal (MP) India. 2 Department of Applied Mathematics and Humanities,S.V. National Institute Of Technology, Surat, Gujarat India. 3 Department of Computer Applications,Maulana Azad National Institute Of Technology, Bhopal (MP) India. Abstract: In this paper, a fuzzy soft set approach is proposed for mining association patterns from a transactional data set.This approach is proposed to capture the uncertain item relationships in data sets and enhance the precision in association rule mining. The transactional data set is represented as soft set using the concept of parameter co-occurrences in the transaction. Thus, each transaction is transformed into soft transaction to generate soft transactional dataset. For the fuzzy parameters present in the transactions, the fuzzy membership is determined and used to transform the soft transactional data set into fuzzy soft transactional dataset. The mining of association patterns is performed on the fuzzy soft transactional data set. The fuzzy soft set approach has been illustrated with the help of an example and experiment on a real world data set. The results of fuzzy soft set approach have been compared with those obtained by traditional, fuzzy set and soft set approaches for mining association patterns. The fuzzy soft set approach gives better picture of association relationship, confidence levels and is helpful in addressing the issues of under-prediction and over-prediction of association patterns. Keywords: Fuzzy Soft set, Soft Set, Fuzzy soft transaction, Fuzzy soft Association pattern I. INTRODUCTION The databases of organizations today consist of large volumes of data piled up due to every day transactions. The managers, scientists, engineers and other decision makers of these organizations are deeply concerned with extracting information and knowledge from these databases for decision making and other applications. A number of data mining techniques are reported in the literature (Atanassov 1994) for extracting information and knowledge from these databases. Association rule mining techniques are widely used by various decision makers of the organizations to explore association patterns in these databases. A good number of algorithms are reported in the literature (Agrawal et al. 1993; Agrawal and Srikant 1994; Mustafa et al. 2006) for mining association rules in databases under deterministic conditions. However the real world data in the fields of engineering, science, technology, biology, medicine and business etc. contains lot of uncertainty, vagueness and impreciseness. The presence of different types of uncertainty in these data sets poses challenges for decision making and mining patterns in these data sets.Various types of theories Eg. theory of Probability, fuzzy sets(Chen and Weng 2009; Zadeh 1965), intuitionitic fuzzy sets(Atanassov 1986), vague sets(Gau and Buehrer 1993) and Rough set (Pawlak 1982) are being employed for dealing with the uncertainties. These theories have been used to deal with different types of uncertainty in data for decision making as well as development of data mining approaches. A number of attempts are reported for development of fuzzy set (Intan 2006), rough set (Guan et al. 2003) and vague set (Lu et al. 2007) approaches for association rule mining .However these rough set, fuzzy set, vague set and probability theories have their inherent difficulties in handling various kinds of uncertainty.Consequently, (Molodtsov 1999) proposed a completely new approach for modeling vagueness and uncertainty called soft set theory which is free from the difficulties present in existing methods was reported in the literature. Some researchers have initiated (Molodtsov 1999; Xiao et al. 2005; Maji et al. 2003)the work with some proposition on soft set operation (Ali et al. 2009) pointed out some new assertions on soft set. With the establishment and development of soft set theory, its wide applications are reported in recent years and are extended to data analysis (Zou and Xiao 2008), decision-making (Maji et al. 2002; Roy and Maji 2007; Cagman and Enginoglu 2010), evaluation(Xiao et al. 2012), medical diagnosis(Borgohain and Das 2010). In soft set theory, membership is decided by adequate parameters, rough set theory employs equivalence classes, whereas fuzzy set theory depends upon grade of membership. Although theories are quite distinct yet deal with vagueness. Joint application of these theories may result in a fruitful way (Lin 1996).Especially on the issue of combination between fuzzy sets, soft sets, rough set, Intuitionistic fuzzy set in these field many researchers have proposed some new concepts on it. Research worker have established mathematical approaches of vagueness such as fuzzy soft set(Maji et al. 2001; Cagman et al. 2010; Çagman et