Churn prediction has received much attention in the last decade. With the evolution of social networks and social network analysis tools in recent years, the consideration of social ties in churn prediction has proven promising. One possibility is to use energy diffusion models to model the spread of influence through a social network. This paper proposes a novel churn prediction diffusion model based on sociometric clique and social status theory. It describes the concept of energy in the diffusion model as an opinion of users, which is transformed to user influence using the derived social status function. Furthermore, a novel diffusion model prediction scheme applicable to a single user or a small subset of users is described: the Targeted User Subset Churn Prediction Scheme. The scheme allows fast churn prediction using limited computing resources. The diffusion model is evaluated on a real dataset of users obtained from the largest Slovenian mobile service provider, using the F-measure and lift curve. The empirical results show a significant improvement in prediction accuracy of the proposed method compared with the basic spreading activation technique (SPA) diffusion model. More specifically, our approach outperforms a basic SPA diffusion model by 116 % in terms of lift in the fifth percentile.U. Droftina (B) · M. Štular

No abstract

In the last years, customer churn prediction has been very high on the agenda of telecommunications service providers. Among customers predicted as churners, highly influential customers deserve special attention, since their churns can also trigger churns of their peers. The aim of this study is to find good predictors of churn influence in a mobile service network. To this end, a procedure for determining the weak ground truth on churn influence is presented and used to determine the churn influence of prepaid customers. The determined scores are used to identify good churn-influence predictors among 74 candidate features. The identified predictors are finally used to build a churn-influence-prediction model. The results show that considerably better churn prediction results can be achieved using the proposed model together with the classical churn-prediction-model than by using the classical churn-prediction model alone. Moreover, the successfully predicted churners by the combined approach also have a greater number of churn followers. A successful retention of the predicted churners could greatly affect churn reduction since it could also prevent the churns of these followers. Key words: Churn prediction, User influence, Social network, Weak ground truth, Churn-influence modelPredvianje odljeva utjecajnih mobilnih pretplatnika korištenjem značajki niske razine. Posljednjih godina, predvianje odljeva korisnika jedna je on važnijih tema meu pružateljima telekomunikacijskih usluga. Meu odlazećim korisnicima, oni najutjecajniji zaslužuju posebnu pažnju, jer njihov odljev može okinuti i odljev sljedbenika. Cilj ovogčlanka je pronalazak dobrih prediktora utjecaja odljeva na mobilne uslužne mreže. U tu svrhu, razvijena je metoda za njihovu identifikaciju meu 74 potencijalna kandidata. Identificirani prediktori su potom korišteni za konačnu izgradnju modela predvianja odljeva korisnika. Znatno bolji rezultati ostvaruju se kada se koristi predloženi model u kombinaciji s klasičnim modelom, nego kada se klasični model koristi zasebno. Štoviše, kombiniranim predvianjem izdvojeni utjecajni korisnici imaju veći broj sljedbenika. Uspješno zadrža-vanje predvienog odljeva moglo bi uvelike utjecati na njegovo smanjenje, pošto bi samim time spriječilo i odljev sljedbenika.Ključne riječi: predvianje odljeva, utjecaj korisnika, društvena mreža, slabi referentni podaci, model utjecaja odljeva

No abstract