Posted on 22 March 2013 by Feri
Mining Food Industry’s Multidimensional Data to Produce Association Rules using Apriori Algorithm as a Basis of Business Strategy in ICOICT 2013
Abstract—The food industry sell a range of product variations. The company want to take advantage to build business strategy from huge information which is stored in data warehouse. In this case, data mining technology needs to be implemented to explore valuable information on transactional data to assess customer’s preferences for products sold as a business strategy.
Information about the way customers buy food products is necessary, this can be done by mapping the transaction data which is described as the pattern of consumer’s taste. The association method using apriori algorithm is used to map customer’s choice.
The challenge is in the data itself, multidimensional data has to be prepared first before the data is fetched to the mining process. Data reduction will be held to handle huge instances and attributes between the data. Research focus on the way we handle data until the rules is built. To reach this goal, three validation levels will be implemented to verify the reliability of the association rules shows by percentage support and confidence.
Keywords—Data Reduction, Apriori, Support, Confidence, Association Rules, Three Validation Levels.
Conclusion according to the paper :
Multidimensional data requires different handling depending what the rules are about to make, for example to build association rules, feature selection should be associated with attributes which wants to build rules. Reduction of association rules consider the presence of other information in the dataset then the reduction should be done systematically consider the linkages between attributes, named briefly as FSA-Red algorithm, which are forwarded to the analysis of the whole dataset, so the end result remains the rule-making association based on the whole dataset available so that no information lost , in this paper three steps of validation were used to analyze whether the association rules which been built is reliable to describe the whole data, in three different conditions : data training after reduction, data training without reduction and data testing.
Many advantage according to the association rules. For example in business point of view, product recommendation is a mature strategy and still effective nowadays, collection of association rules can be used to build such strategy for example the way product will be displayed or sell few products as a packages which is easier to reach to increase sales[5] .
References :
[1] Agrawal, R., Srikant, R., Fast Algorithms for Mining Association Rules. IBM Almaden Research Center 650 Harry Road, San Jose, CA 95120. Proceedings of the 20th VLDB Conference Santiago, Chile. 1994.
[2] Campos, Marcos., Stengard,Peter., Milenova,Boriana., Data Centric Automated data Mining. Proceedings of the Fourth International Conference on Machine Learning and Applications (ICMLA’05). IEEE. 2005.
[3] D.N, Goswami., Anshu, Chaturvedi., C.S, Raghuvanshi., D.N, Goswami. et. al.An Algorithm for Frequent Pattern Mining Based On Apriori. (IJCSE) International Journal on Computer Science and Engineering Vol. 02, No. 04, 2010, 942-947.
[4] Das, Gautama., Lin, King-Ip., Mannila, Heikki,. Renganathan, Gopal., Smyth,Padhraic., Rule discovery from time series. American Association for Artificial Intelligence.1998.
[5] Davenport, Thomas., Realizing the Potential of Retail Analytics- Plenty of Food for Those with the Appetite. Babson Working knowledge research report. 2009.
[6] Euler,Timm., Modeling preparation for data mining processes. Journal of Telecommunications and Information Technolog. 2006.
[7] Farajian, Mohammad Ali., Mohammadi, Shahriar., Mining the Banking Customer Behaviour Using Clustering and Association Rules Methods, International Journal of Industrial Engineering and Production Research.Vol 21, Number 4 pp. 239-245.2010.
[8] Haery,A., Salmasi,N., Modarres Yazdi,M., Iranmanesh,H., Application of Association Rule Mining in Supplier Selection Criteria. World Academy of Science, Engineering and Technology 40 2008.
[9] Han,Jiawei., Kamber, Micheline., Data Mining:Concepts and Techniques. Morgan Kaufmann Publishers 2006 page.: 4-37 , page : 227-260.
[10] Orlando,S., Palmerini,P., Perego,R., Silvestri,F., Adaptive and resource-aware mining of frequent sets. Proceedings of the IEEE International Conference on Data Mining. IEEE Computer Society. 2002.
[11] Parthasarathy, S., Efficient progressive sampling for association rules. IEEE International Conference on Data Mining, pp.: 354- 361, 2002.
[12] Witten,Ian., Frank, Eibe., Data Mining – Practical Machine Learning Tool and techniques. Morgan Kaufmann Publishers 2005 page :6 ; page: 27 paragraph 1 ; page : 112-118, page 47-86.
External link point to the ICOICT 2013 conference :
No comments:
Post a Comment
Note: Only a member of this blog may post a comment.