Using Fingerprint Recognition Technology Mining Telecommunication Reentry Net Customer

How to control the problem of repeat network access is a headache for the mobile network operators. It has attracted gem attention, which can lead to the increase of operation cost and decrease of business earning because of the waste of card number resource arrearage.In this paper, combining with the actual production, defining customer's phone call records characteristics as the calling fingerprint, through two-step classification, the author divides new access target customers into re-access customers and non-reaccess customers. In the first step, comparing the new customer's call fingerprint to the fingerprint of the history number, according to the degree of similarity, the list of re-access suspects among the customers will be identified. Combining IMEI information with customer's identification information, class the new access customers of lower similarity as non-reaccess customers, and which of higher similarity as re-access customers, and which of medium similarity as the re-access number collection to be determined, then move on the next comparison. In the second step, adopting sorting algorithm to further discover the re-access customers, taking the re-access and non-reaccess number collection of the first step as a training collection, make comparison and analysis of a variety of sorting algorithms, present comprehensive consideration from the algorithm execution speed, scalability and intelligibility of output results. In light of the main considerations of the forecasted accuracy and calculating speed, using determination rules generated from the improved ID3 algorithm, class the number collection of the first step which are temporarily unable to be determined, and the numbers are relegated to the re-access and non-reaccess number collection. Finally, the target customers will be divided into re-access and non-reaccess customers.The re-access customer identification algorithm in this paper has been implemented in The Business Analysis Support System for China Mobile. The results show that this algorithm has effectively identify the re-access customers, developed re-access customer channels analysis, re-access customer marketing policy analysis, specific applications of re-access customers recharge and arrearage analysis, has acted as a perfect date support for China Mobile in reducing operational risk as the waste of number resources, arrearage rate, bulk re-access customers caused by tariff adjustment.