SIMBOX Identification Using K-Nearest Neighbor Based On Spectrum Analyzer

Abstract

Telecommunication Service Provider should deal with illegal players (grey operators) who do not have permission to conduct international voice service. These illegal players perform their activities by passing international incoming traffic using Simbox devices. To identify visual simbox usage is very difficult and less reliable, therefore by using spectrum analyzer and K-Nearest Neighbor (K-NN) method is one way to identify simbox usage. The attributes used in the identification process are Location / Document, Strong Frequency Signal, and by applying K-NN algorithm based on proximity of training data with data testing. The determination of this attribute is based on GSM DCS 1800 MHz uplink frequency measurement in Cilacap and Banyumas area. The identification process was conducted on six frequencies points on 18 data with the largest signal strength as training data. Moreover, the signal strength data testing  by using 32 data gives result 81.25% accuracy. The results of K-NN algorithm calculations can be implemented to identify the use of simbox, hence it can be used as a reference for mobile operators to identify simbox usage in other areas.Telecommunication Service Provider should deal with illegal players (grey operators) who do not have permission to conduct international voice service. These illegal players perform their activities by passing international incoming traffic using Simbox devices. To identify visual simbox usage is very difficult and less reliable, therefore by using spectrum analyzer and K-Nearest Neighbor (K-NN) method is one way to identify simbox usage. The attributes used in the identification process are Location / Document, Strong Frequency Signal, and by applying K-NN algorithm based on proximity of training data with data testing. The determination of this attribute is based on GSM DCS 1800 MHz uplink frequency measurement in Cilacap and Banyumas area. The identification process was conducted on six frequencies points on 18 data with the largest signal strength as training data. Moreover, the signal strength data testing  by using 32 data gives result 81.25% accuracy. The results of K-NN algorithm calculations can be implemented to identify the use of simbox, hence it can be used as a reference for mobile operators to identify simbox usage in other areas.