Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
International Journal of Computer Science Engineering and Information Technology Research (IJCSEITR) ISSN 2249-6831 Vol. 3, Issue 3, Aug 2013, 65-78 © TJPRC Pvt. Ltd. EFFICIENT APPROACH FOR EXTRACTING FREQUENT PATTERN AND ASSOCIATION RULES WITH PERIODIC CONSTRAINTS VARSHA MASHORIA1 & ANJU SINGH2 1 Department of Computer Science, Barkatullah University, Institute of Technology, Madhya Pradesh, India 2 Department of Information Technology, Barkatullah University, Institute of Technology, Madhya Pradesh, India ABSTRACT Mining of association rule from huge amounts of data in the transaction database plays a vital role in data mining, but as technology develops it provide us with many techniques for mining rules. For improving the performance and capability of the association rule mining task, we use different techniques .Constraint based mining is one of them, with the help of this technique the user can think about closely on mining only interested association rule instead of complete set of association rule.. In this paper we propose an approach basically for mining association rules with periodic constraint, which consist of three phases .In first phase it generate the frequent itemset then in the second phase we exploit properties of periodic constraint and in the last phase it generate all the patterns or rules. KEYWORDS: Association Rule Mining, Conditional Database, Constraint Based Mining, Frequent Itemsets, Minimum Support, Periodic Constraints INTRODUCTION Data mining is the analysis step of the knowledge Discovery from data (KDD) process. Mining association rules is a very substantial problem in the data mining field [4]. Which is defined as the non trival eradication of earlier unknown and potentially useful facts from data in database[27].Discovering association rules between items in a huge transactional database plays an vital role in data mining research areas. An association rule is an suggestion of the form A and B are frequent item sets in a transaction database and A∩B = Ǿ. The rule A B, where A B can be explained as “if item set A occurs in a transaction, then item set B will also likely occur in the same transaction”. By such type of facts, market personnel can place item sets A and B within close immediacy, which may encourage the sale of these items together and develop discount strategies based on such association/correlation found in the data. Therefore, association rule mining has received a lot of attention. With the growth of data mining techniques, quite a few researchers have worked on alternative patterns. In many (but not all) situations, we only care about association rules or idea involving sets of items that appear frequently in baskets. For example, any one cannot run a good marketing strategy involving items that no one buys anyway. Thus, much data mining starts with the assumption that we only care about sets of items with high support; i.e., they appear together in many baskets. We then find association rules or generator only involving a high-support set of items must appear in at least a certain percent of the baskets, called the support threshold [25]. Frequent pattern mining generate huge amount of frequent itemsets and rules . Frequent patterns are itemsets, consecutiveness, or substructures that appear in a data set with frequency no less than a user-specified threshold. For example, a set of items, such as milk and bread that appear frequently together in a transaction data set is a frequent itemset. A concatenation, such as buying first a PC, then a digital camera, and then a memory card, if it occurs frequently in a shopping history database, is a (frequent) sequential pattern. A substructure can refer to different structural forms, such as 66 Varsha Mashoria & Anju Singh sub graphs, sub trees, or sub lattices, which may be combined with itemsets or consecutiveness. If a substructure occurs frequently in a graph database, it is called a (frequent) structural pattern. Finding frequent patterns plays an essential role in mining associations, correlations, and many other affecting relationships among data. Moreover, it helps in data indexing, classification, clustering, and other data mining tasks as well. Thus, frequent pattern mining has become an important data mining task and a focused theme in data mining research [24]. Frequent pattern mining was first proposed by [3] Many constraint-based frequent itemset discovery techniques have been proposed for various constraint models which not only reduce the effectiveness but also reduce the efficiency of mining and the pattern which is generated from the techniques are repetitious in nature..It plays an essential role in mining associations. Item constraints in frequent itemset mining were first discussed in [22], which considered the problem of integrating constraints that are Boolean expressions over the presence or absence of items into the association rule discovery algorithms. [23] Introduced two affecting classes of itemset constraints: anti-monotonicity and succinctness, and proposed a mining algorithm for handling constraints belonging to these classes within the Apriori framework, called CAP. [26] Developed a new constraint-based frequent itemset discovery method, called CFG, which pushed a constraint into the FPgrowth method. In [21], the third class of constraint, monotonicity, was introduced in the context of mining correlated sets. In [20], more classes of constraints were introduced: convertible constraints, and methods which enable these classes of constraints to be pushed deep inside the FPgrowth algorithm for frequent itemset mining, were developed. [18] Proposed the Dual Miner algorithm and used both monotone and anti-monotone constraints to prune the search space. [19] Presented a method to mine itemsets with restrictions on their variance. [15] Introduced a new class of block constraints that determined the acceptation of an itemset pattern In most of the previous constraint based association rule methods the goal is to discover all the patterns whose frequency in the dataset exceeds a user-specified threshold. However, very often users want to restrict the set of patterns to be detected by adding extra constraints on the structure of patterns. Data mining systems should be able to exploit such constraints to speed up the mining process. Techniques applicable to constraint-driven pattern discovery can be classified into the following groups: ' post-processing (filtering out patterns that do not satisfy user specified pattern constraints after the actual discovery process); ' pattern filtering (unification of pattern constraints into the actual mining process in order to generate only patterns satisfying the constraints); ' dataset filtering (restricting the source dataset to objects that can possibly contain patterns that satisfy pattern constraints)[12]. Wojciechowski and Zakrzewicz [17] focus on improving the efficiency of constraint- based frequent pattern mining by using dataset filtering techniques. Dataset filtering conceptually transforms a given data mining task into an equivalent one operating on a smaller dataset. Tien Dung Do et al [13] proposed a specific type of constraints called category-based as well as the associated algorithm for constrained rule mining based on Apriori. The Category-based Apriori algorithm reduces the computational complexity of the mining process by get around most of the subsets of the final itemsets. An experiment has been conducted to show the efficiency of the proposed technique. In this proposed algorithms we basically used periodic functions as constraints in rule mining techniques. The periodic constraints functions have infinite number of solution, but usually there are constrained periodic value varied as min, max interval for the validation of minimum support and confidence parameter for rule generation. The process of rule mining based on periodic function first abstract the variable, and then eliminates all solution that fall outside at axiom constraints. The remaining of this paper is organized as follows. In Section 2, we discuss the problem statement. In Section 3, we describe Periodic constraint. Next, In section 4 we describe the related work and work then we describe the main idea in Section 5, in which we describe Future work. Section 6 consists of a series of experiment or performance evaluation. Efficient Approach for Extracting Frequent Pattern and Association Rules with Periodic Constraints 67 Conclusions are presented in Section 7.Section 8 represent the future work. PROBLEM STATEMENT In the following, we give a formal statement of the problem for which we develop an algorithm for mining all sophisticated association rules with the help of periodic constraints, so that only sophisticated rule can be generated which fulfill all the given parameters. The algorithm parameter specifying the minimum confidence bound is known as minconf and the minimum support bound minsup. A huge number of Association rules can be found from a transactional dataset. To find the interesting Association rules in a transactional dataset, we must define a specified minimum support (called minsup) and specified minimum confidence (called minconf). The itemset Y I is called frequent itemset if sup (X) => minsup. It is known that a subset of frequent itemset is a frequent itemset any superset of infrequent itemset is not a frequent itemset. Finally, the Association rule X→Y holds if conf (X→Y) >= minconf. The Association rules are said to be strong if it meets the minimum confidence threshold. However, while Association rules provide means to discover many interesting Associations. Example show how to calculate Association rule from the transaction dataset is explained later. Association rules are widely used in various areas such as telecommunication networks, market and risk management, inventory control etc. Various Association mining techniques and algorithms will be briefly introduced and compared later. In many cases, the algorithms generate an extremely large number of Association rules, often in thousands or even millions. Further, the Association rules are sometimes very large. It is nearly impossible for the end users to comprehend or validate such large number of complex Association rules, thereby limiting the usefulness of the Data Mining results. Several strategies have been proposed to reduce the number of Association rules, such as generating only “interesting” rules, generating only “non redundant” rules, or generating only those rules satisfying certain other criteria such as coverage, leverage, lift or strength. In general, Association rule mining can be viewed as a two-step process: Step 1. Find all frequent itemsets: By definition, each of these itemsets will occur at least as frequently as a predetermined minimum support count, min sup. Step 2. Generate strong Association rules from the frequent itemsets: By definition, these rules must satisfy minimum support and minimum confidence. There are many algorithms is developed for finding frequent patterns and on the basis of frequent patterns, [29] Association rule is generated. With the help of constraint we solve this type of problem. Constraints do two things: 1) They limit where the algorithm can look; and 2) they give hints about where to look. PERIODIC CONSTRAINTS In its most general development of entity, the task of data mining is vastly undefined. To make the task more absolute, we first have to specify the type of patterns considered such as frequent patterns, a clustering, a predictive model or other invariability in the data. But the detected patterns may not be novel or actionable in fields where domain expertise already exists or users have strong anticipation. We then have to specify what conditions the patterns have to satisfy in order to consider them as solutions to the data mining task at hand. The conditions that a pattern has to satisfy can be elegantly specified as constraints, stated explicitly and under direct control of the user/data miner. Constraints play vital role in data mining as the use of constraints enables more efficient data mining and focuses the search for patterns on patterns likely to be of interest to the end user. The ability to express and exploit constraints allows the data miner to inject knowledge into the process of data mining and 68 Varsha Mashoria & Anju Singh knowledge discovery. The different type of constraints is defined for sets [23] that can also be applied in structured data mining. The structured of data can be ignored and the constraints can be applied as a post processing step of the mining algorithm. An instance of the periodic constraints is a finite set of “generating” constraints over a large, structured variable set, for instance, the set of lattice points in k-dimensional space. A correspondent full set of constraints is obtained by repeating the generating set periodically, and the problem is to decide whether or not the full set of constraints has a satisfying task. This model is natural for studying large, possibly infinite constraint networks consisting of constraints obeying a high degree of regularity or symmetry. Such constraint networks arise naturally in many domains, such as scheduling, planning, and hardware design. So that our results may enjoy maximal applicability. Definition A function f(x) is said to be periodic (or when emphasizing the presence of a single periodic [figure 1(a)] instead of multiple periods singly periodic) with period p if f(x) = f(x + np) for n = 1, 2, 3……as. for e.g. the sin function sinx illustrated below is periodic with least period (often simple the period) 2Π (as well as will period = 2Π, 4Π, 6Π…..etc).The constraint function f(x) = 0 is periodic with any periodic with any period R for all non zero real number R1 so there is no concepts analogeous to the least period for constant function the following definition summarizes the names given to periodic function based on the number of independent periods they posses. Let f be a function for item (I1,I2,..........In) there axiom interval (min , max), We define the periodic function in two mode even and odd. Even periodic [figure1 (b)] f= (0<min and min<max) f (f1,f2,.........,fn) , n<0 and -min<n Odd periodic [figure1 (c)] f= (-<max and max<min) f (f1,f2,.........,fn) , n<0 and -max<n Figure 1(a): Figure Showing the Basic Concept of Periodic Constraint Single Periodic Constraints A function processing a single period in the complex plane is said to be single periodic often simply periodic. This function includes the trigonometric function. Efficient Approach for Extracting Frequent Pattern and Association Rules with Periodic Constraints 69 Doubly Periodic Function A function f(x) is said to be doubly periodic, if it has 2 period W1 and W2 whose ratio W2/W1 is not real. A doubly periodic function that is analytic (except at poles) and that has no singularities other than poles the finite plane is called an ellipse function. The period W1 and W2 plays the same part in the theory of elliptic function as does the single period the case of the trigonometric function. Figure 1(b): Even Function as Even Constraint Figure 1(c): Odd Function as Odd Constraint Triply Periodic Function A function having 3 distinct periods is known as triply period function. Jacobi 1835 proved that a single – valued univariate function cannot have more than 2 distinct periods. Thus showing that elliptic function is the most general multiply periodic single – valued function possible as a single variable. Analytically, f is an even function if its domain contains the point –x whenever it contains x, and if f (-x) = f (x) for each x in the domain of f. The function f is an odd function if its domain contains the point –x whenever it contains x, and if f (-x) = - f (x) for each x in the domain of f. RELATED WORK Constrained data mining has been said to be the “best division of labour,” where the computer does the number crunching and the human provides the focus of attention and direction of the search by providing search constraints.[9] Constraints do two things: 1) They limit where the algorithm can look; and 2) they give hints about where to look. [5] As a constraint is a guide to direct the search, combining knowledge with inductive logic programming is a type of constraint, and that knowledge directs the search and limits the results. This combination is extremely effective [16]. If every possible pattern is selected and the constraints tested afterwards, then the search space becomes large and the time required to perform this becomes excessive. [2] The constraints must be in place during the search. They can be as simple as thresholds on rule quality measure support or confidence, or more complicated logic to formulate various conditions. [11] In mining with a structured query language (SQL), the constraint can be a predicate for association rules. [17] In this case, the rule has a constraint limiting which records to select. This can either be the total job or produce data for a next stage of refinement. For example, in a large database of bank transactions, one could specify only records of ACH transactions that occurred during the first half of this year. This reduces the search space for the next process. A typical search would be: select * where year = 2006 and where month < 7 It might be necessary that two certain types always cluster together (must-link), or the opposite, that they may never be in the same cluster (cannot-link).[6] In clustering (except fuzzy clustering), elements either are or are not in the same cluster. [2] Application of this to the above example could further require that the 70 Varsha Mashoria & Anju Singh transactions must have occurred on the first business day of the week (must-link), even further attenuating the dataset. It could be even further restricted by adding a cannot-link rule such as not including a national holiday. In the U.S.A., this rule would reduce the search space by a little over 10 percent. The rule would be similar to: select * where day = Monday and day <8 and where day \=holiday If mining with a decision tree, pruning is an effective way of applying constraints. This has the effect of pruning the clustering dendogram (clustering tree). If none of the elements on the branch meet the constraints, then the entire branch can be pruned [2] In Ravi and Davidson's study of image location for a robot, the savings from pruning were between 50 percent and 80 percent. There was also a typical improvement of a 15 percent reduction in distortion in the clusters, and the class label purity improved. Applying this to the banking example, any branch that had a Monday national holiday could be deleted. This would save about five weeks a year, or about 10 percent. Mining a large data set can be time consuming, and without constraints, the process could generate sets or rules that are invalid or redundant. Some methods, for example association rule mining, are effective, but can be extremely time consuming for large data sets. As the set grows in size, the processing time grows exponentially. In other situations, without guidance via constraints, the data mining process might find morsels that have no relevance to the topic or are trivial and hence worthless However, frequent itemset mining algorithms often generate a large number of frequent itemsets and rules, which reduce not only the efficiency but also the effectiveness of the mining algorithms since only the subset of the complete frequent itemsets and association rules is of interest to users, and users need additional post-processing to filter through a large number of mined rules to find the useful ones. Constraint-based mining enables users to provide restraints to search for those useful ones. It concentration itemsets that are interesting from a user’s point of view, rather than wasting time on discovering itemsets the user has not asked for. Many constraint-based frequent itemset discovery techniques have been proposed recently for various constraint models. Item constraints in frequent itemset mining were first discussed in [24], which considered the problem of integrating constraints that are Boolean expressions over the presence or absence of items into the association rule discovery algorithms.[17] introduced two interesting classes of itemset constraints: antimonotonicity and succinctness, and proposed a mining algorithm for handling constraints belonging to these classes within the Apriori Framework, called CAP.[18] developed a new constraint-based frequent itemset discovery method, called CFG, which pushed a constraint into the FPgrowth method. In [8], the third class of constraint, monotonicity, was introduced in the context of mining correlated sets. In [19], more classes of constraints were introduced: convertible constraints, and methods which enable these classes of constraints to be pushed deep inside the FP-growth algorithm for frequent itemset mining, were developed.[4] proposed the DualMiner algorithm and used both monotone and anti-monotone constraints to prune the search space.[14] presented a method to mine itemsets with restrictions on their variance.[7] introduced a new class of block constraints that determined the significance of an itemset pattern. In most of previous constraint-based frequent itemset mining methods, an item in a transaction is identified by its item ID and characterized only by a single attribute value (such as price). In some circumstances, users may want to keep records of items with more than one attribute. So we in this paper present a method of association rule mining with periodic constraint, with the help of which we mine the association rule which is of user interest. So that all the redundant and invalid rules which does not satisfy the user specified condition. In comparison with the Apriori algorithm [1], the FP-growth algorithm [10], is more efficient to mine frequent itemsets. Furthermore, FP-growth-based methods also achieve performance gain in constraint-based mining by the employment of conditional databases. More properties of constraints can be pushed into FP-growth-based methods than Apriori-like methods. Therefore, in this paper, we choose the FP-growth method as the basic approach in our model and develop the algorithms to mine frequent itemsets. Efficient Approach for Extracting Frequent Pattern and Association Rules with Periodic Constraints 71 PROPOSED WORK AND PROPOSED MINING ALGORITHM Association Rule Mining is key player of pattern analysis of affecting data. The size of data is big and the generation of pattern and rule is huge. The generated huge pattern and rule are no meaningful. So now in these day constraints based association rule mining algorithm is applied. The constraints factor reduces the unnecessary rule set in process of rule generation. Some constraints based method such as monotonic and non monotonic or anti monotonic. These are very effective for certain valid pattern generation but lacking some constraints factor such as continuity factor generates some invalid rule. For remedies of these processes we used periodic functions as constraints in rule mining techniques. The periodic constraints functions have infinite number of solution, but usually there are constrained periodic value varied as min, max interval for the validation of minimum support and confidence parameter for rule generation. The process of rule mining based on periodic function first abstract the variable, and then eliminates all solution that fall outside at axiom constraints. Figure 2: Data Flow Diagram Now in mining algorithm by fulfilling all condition put by periodic constraint all frequent items are generated. In this work the proposed algorithm deal with 2 periodic constraints that is odd constraints and even constraints. A periodic constraints C is convertible into odd constraints and even constraints. The set of constraints are not statically the rule mining process the set of rule generation is Ø. After generating frequent item we apply FP-Growth algorithm so that all rules can be generated by it. The dataflow diagram in Figure 2 is a solution to this problem. There is one decision symbol in the data flow diagram, which checks for constraint. If the frequent item generated does not fulfill the constraints condition it will go back to second step and if the frequent item generated fulfill the condition of constraint it will proceeded to next step. In next step the item generated by the decision symbol is divided into odd and even. after this we find the set of odd and even items the condition is check for Ø. After this we use the FP-Growth algorithm to generate the rules. Scan each item from the transaction table. Priori compute even and odd set of items according to given minimum and maximum condition. Mapping statically all constraint itemset and their support. 72 Varsha Mashoria & Anju Singh Eliminate the value of outside constraints. Sort the even and odd itensets. Check for (i=1; Eoij≠0;i++) Eoij = FPgrowth(Eoij) generate new coordinate. Delete all itemsets generate from the outage variable in set Eoij. return frequent rules. RESULTS AND DISCUSSIONS For evaluation purpose, we have conducted several experiments. We select the well-known IBM synthetic database generated by the generator in [28]. This choice gives us the flexibility of controlling the size and dimension of the database. All of experiments are performed on Pentium IV processor with a 3.4 GHz processor and 2 GB memory. The program is developed in MATLAB 7.8.0. In this paper, the runtime includes both CPU time and I/O time. In all of the following experiments, we limit the length of association rules. Since our algorithms can be applied to mine association rules among frequent items with bounded length, we compare our schemes with the most influential algorithms for association rule mining, which are the Apriori algorithm and FP-Tree. . The parameter we used for compare our algorithm are maximum (max) value, minimum (min) value, minimum support (minsup) and minimum confidence (minconf). The calculated result are in the form of number of rule generated and the execution time it takes to generate these rules with the given parameter. The values of parameter are changed every time and the values are recorded in different tables this figure shows the comparative result between Apriori, FP-Tree and our proposed algorithm. All the values generated from this can be plotted on graph. Table 1(a): Shows the Calculated Result of Generated Rules with Apriori, FP-Tree and Our Proposed Approach Figure 3(a): Shows the Comparission Graph between Number of Rules Generated by Using Different Dataset Chess, Wine, Mushroom and with Different Methods Efficient Approach for Extracting Frequent Pattern and Association Rules with Periodic Constraints Table 1(b): Shows the Calculated Result of Execution Time Taken by these Approach to Generate the Rules Figure 3(b): Shows the Comparission Graph between Execution Time Taken by Different Dataset (Chess, Wine, Mushroom.etc) and with Different Methods Table 2(a): Shows the Calculated Result of Generated Rules Apriori, FP-Tree and Our Proposed Approach Figure 4(a): Shows the Comparission Graph between Number with of Rules Generated by Using Different Dataset (Chess, Wine, Mushroom etc) and with Different Methods 73 74 Varsha Mashoria & Anju Singh Table 2(b): Shows the Calculated Result of Execution Time Taken by these Approach to Generate the Rules Figure 4(b): Shows the Comparission Graph between Execution Time Taken by Different Dataset (Chess, Wine, Mushroom.etc) and with Different Methods Table 3(a): Shows the Calculated Result of Generated Rules with Apriori, FP-Tree and our Proposed Approach Figure 5(a): Shows the Comparission Graph between Number of Rules Generated by Using Different Dataset Chess, Wine, Mushroom etc) and with Different Methods Efficient Approach for Extracting Frequent Pattern and Association Rules with Periodic Constraints 75 Table 3(b): Shows the Calculated Result of Execution Time Taken by these Approach to Generate the Rules Figure 5(b): Shows the Compression Graph between Execution Time Taken by Different Dataset (Chess, Wine, Mushroom.etc) and with Different Methods CONCLUSIONS In this paper we proposed a novel method for pattern mining based on periodic constraints. As we know that constraints are the condition that a pattern has to satisfied. With the help of periodic constraint we generate all valid (user defined) rules and all the unnecessary rules which does not satisfy user level are reduced or deleted. In this paper we use periodic functions as constraints in rule mining techniques, Usually there are constrained periodic value varied as min, max interval for the validation of minimum support and confidence parameter for rule generation. The process of rule mining based on periodic function first abstract the variable, and then eliminates all solution that fall outside at axiom constraints. The set of constraint are not statically, the rule mining process the set of rule generation is Ø. Our experimental result shows that our algorithm attains very good mining efficiencies on various input datasets. FUTURE WORK The limitation of this algorithm is that the execution time increase as our dataset increases. So these investigations have identified some interesting directions for future research: In the case of minimum, maximum condition occurred a data over fitting the situation of rules generation is all positive and all negative. We also hide the sensitive pattern using minimum, maximum condition ( Data preserving ). REFERENCES 1. Agrawal, R., Srikant, R., " Fast algorithms for mining association rules", In: Proceedings of International Conference on Very Large Data Bases, 1994 , pp. 487–499. 76 Varsha Mashoria & Anju Singh 2. Boulicaut, J-P., & Jeudy, B., "Constraint-Based Data Mining", The Data Mining and Knowledge Discovery Handbook 2005, Maimon, O. & Rokach, L. Eds Springer-Verlag , 2005 ,399-416. 3. Brin, S., Motwani, R. and Silverstein, C., "Beyond Market Baskets: Generalizing Association Rules to Correlations" Proc. ACM SIGMOD Conf., pp. 265-276, May 1997. 4. Bucila, C., Gehrke, J., Kifer, D., White, W., "DualMiner: a dualpruning algorithm for itemsets with constraints", In: Proceedings of ACM International Conference on Knowledge Discovery and Data Mining, pp. 42–51,2002. 5. Davidson, I., & Ravi, S.," Agglomerative Hierarchical Clustering with Constraints: Theoretical and Empirical Results", Proceedings of PKDD 2005 9th European Conference on Principles and Practice of Knowledge Discovery in Database, Jorge, A., P., & Gama, J, Eds. Springer, 2005 , 59-70. 6. Davidson, I., & Ravi, S.,"Clustering Under Constraints: Feasibility Issues and the k-Means Algorithm", Papers Presented in the 2005 SIAM International Data Mining Conference, 2005 . 7. Gade, K., Wang, J., Karypis, G., 2004. , "Efficient closed pattern mining in the presence of tough block constraints", In: Proceedings of ACM International Conference on Knowledge Discovery and Data Mining, pp. 138–147. 8. Grahne, G., Lakshmanan, L.V.S., Wang, X., "Efficient mining of constrained correlated sets", In: Proceedings of IEEE International Conference on Data Engineering, pp. 512–521,2002. 9. HanJ., Dong.G., Yin.Y., "Efficient mining of constrained correlated sets", In: Proceedings of IEEE International Conference on Data Engineering . 2000, pp.512-521. 10. Han, J., Pei, J., Yin, Y., " Mining frequent patterns without candidate generation", In: Proceedings of ACMSIGMOD, 2000 , pp. 1–12. 11. Hipp, J., & Guntzer, U. (). ,"Is Pushing Constraints Deeply Into the Mining Algorithm Really What We Want?", An Alternative Approach to Association Rule Mining, SIGKDD, 4(1), 50-55. 12. IBM Almaden.quest synthetic data generation code. http://www.almaden.ibm.com/cs/quest/syndata.html. 13. Jiawei Han · Hong Cheng · Dong Xin · Xifeng Yan, "Frequent pattern mining: current status and future directions" ,Received: 22 June 2006 / Accepted: 8 November 2006 / Published online: 27 January 2007 Springer Science+Business Media, LLC 2007. 14. Kifer, D., Gehrke, J., Bucila, C., White, W., 2003., "How to quickly find a witness", In: Proceedings of ACM SIGMOD-SIGACTSIGART Symposium on Principles of Database Systems, pp. 272–283. 15. M.S. Chen, J. Han, P.S. Yu, "Data mining: An overview from database perspective" , IEEE Transactions on Knowledge and Data Engineering vol. 8 (1996) 866–883,J. Han, M. Kamber, Data Mining Concepts and Techniques, second ed., Morgan Kaufmann, 2005. 16. Muggleton, S., "Scientific Knowledge Discovery Using Inductive Logic Programming", Communications of the ACM, 42(11), November 1999, 42-46. 17. Ng, R., Lakshmanan, L.V.S., Han, J., Pang, A., " Exploratory mining and pruning optimizations of constrained association rules", In: Proceedings of ACM-SIGMOD, pp. 13–24,1998. 77 Efficient Approach for Extracting Frequent Pattern and Association Rules with Periodic Constraints 18. Pei, J., Han, J., " Can we push more constraints into frequent pattern mining?", In: Proceedings of ACM International Conference on Knowledge Discovery and Data Mining, 350–354,2000. 19. Pei, J., Han, J., Lakshmanan, L.V.S., " Mining frequent itemsets with convertible constraints", In: Proceedings of IEEE International Conference on Data Engineering, pp. 433–442,2001. 20. R. Agrawal, T. Imielinsi and A. Swami, “Mining Association Rules between Sets of Items in Very Large Databases”, Prof. ACM SIGMOD Conf Management of Data, pp. 207-216, 1993. 21. Savasere, A., Omiecinski, E., Navathe, S.,"Mining for strong negative associations in a large database of customer transactions". In: Proc. of ICDE. 494–502, (1998). 22. Schuster, A. and Wolff, R., " Communication-efficient distributed mining of association rules" , in `Proc. of the 2001 ACM SIGMOD Int'l. Conference on Management of Data', Santa Barbara, California, pp. 473-484,2001. 23. Sotiris Kotsiantis, Dimitris Kanellopoulos ,"Association Rules Mining: A Recent Overview", Educational Software Development Laboratory Department of Mathematics, University of Patras, Greece [email protected], [email protected] GESTS International Transactions on Computer Science and Engineering, Vol.32 (1), pp. 71-82,2006. 24. Srikant, R., Vu, Q., Agrawal, R., " Mining association rules with item constraints", In: Proceedings of ACM International Conference on Knowledge Discovery and Data Mining, pp. 67–73,1997. 25. Teng, W., Hsieh, M., Chen, M.: On the mining of substitution rules for statistically dependent items. In: Proc. of ICDM. 442–449, (2002) . 26. Tien Dung Do, Siu Cheung Hui, Alvis Fong, "Mining Frequent Itemsets with Category-Based Constraints", Lecture Notes in Computer Science, Volume 2843, pp. 76 - 86,2003. 27. Wojciechowski, M., Zakrzewicz, M., "Dataset Filtering Techniques in Constraint-Based Frequent Pattern Mining", Lecture Notes in Computer Science, Volume 2447, pp.77-83,2002. 28. Wu, X., Zhang, C., Zhang, S.: Efficient Mining of Both Positive and Negative Association Rules, ACM Transactions on Facts Systems, Vol. 22, No. 3, Pages 381–405, July 2004. 29. Yin-Ling Cheung , Ada Wai-Chee Fu, "This paper appears in: Knowledge and Data Engineering", IEEE Transactions on Date of Publication: Dept. of Comput. Sci., Hong Kong Univ., Shatin, China Volume: 16, Issue: 9 Page(s): 1052 - 1069 Product Type: Journals & Magazines, Sept. 2004. 30. Yuan, X., Buckles, B., Yuan, Z., Zhang, J., "Mining negative association rules" In: Proc. Of ISCC, 623–629, 2002.