AN ORDERED FUZZY RULE INDUCTION BASED CHURN MINING FOR TELECOM INDUSTRY

N Kamalraj¹, Dr A Malathi²

¹Dr.SNS Rajalakshmi College of Arts & Science, Coimbatore, India
²Government Arts College, Coimbatore, India

Abstract

Distributed Strategies using outlier mining can result in vast time savings and communication cost while predicting churns in large data sets. However, rule based on the data mining procedure is not obtained with good solutions. Classification Rules based on time series using polynomial modeling was proposed to reduce the churn. But comprehensibility rule set was not addressed while predicting the churns (i.e., faults). In this paper, we propose an Ordered Fuzzy Rule Induction based Churn Mining (OFRI-CM) model to perform the rule mining operation on the clustered churn. The OFRI-CM model develops a system based on the rule that mine the churn from higher order to lower ones. A Fuzzy Data Mining model is first constructed to attain better solution using soft boundaries that perform aggregation operators for combining fuzzy rules, aiming at reducing the true positive rate of churn being detected. Then, we propose a divide and conquer strategy, based on the new sub group pattern of churns (i.e., multiple churns). The final best churns are predicted using divide and conquer strategy that recursively breakdown the churns into two or more churns and are finally aggregated to perform high region of search space with minimal processing time. This process is repeated until a churn mining criterion with comprehensibility rule set is generated. Finally, we present Churn Mining-based Greedy algorithm to improve the accuracy of comprehensibility rule set being generated. Experiments are conducted on Ericson GSM and Nokia Siemens GSM systems and the results show that the Churn Mining-based Greedy algorithm is efficient and that its average processing time of rules scales quite well for an increasing number of subscribers.

Keywords

• Classification Rules, Churn Mining, Fuzzy Data Mining model, Divide and Conquer strategy

Author's Profile

Author profile can be generated and linked through our partners World Book of Researchers. To include your profile online Click Here. After it is approved, please email to edlib @ asdf.res.in to create a link with all the papers.

• N Kamalraj : Profile
• Dr A Malathi : Profile

e-AID

Cite this Article as Follows

N Kamalraj, Dr A Malathi. "AN ORDERED FUZZY RULE INDUCTION BASED CHURN MINING FOR TELECOM INDUSTRY." International Conference on Interdisciplinary Research in Electronics and Instrumentation Engineering (2015): 18-25. Print.