Explaining and Predicting the Loss of Customers for an RSS Feed ReaderApplication

Detta är en Kandidat-uppsats från KTH/Matematisk statistik

Författare: Alexandra Benckert; Danilo Duarte Dos Reis; [2021]

Nyckelord: ;

Sammanfattning: This thesis within mathematical statistics aimed to investigate the most significant factors in predicting and explaining the cancellation of a customer's subscription for an RSS feed reader. The applied method was logistic regression analysis and the thesis was done in cooperation with the Swedish RSS feed reader company Feeder. The data set was obtained from Feeder's database and contained variables describing user behavior, payments, setting, and subscription type. Parts of the obtained data were originally grouped, but additional effort was required to find other relevant groupings of the variables to obtain a reasonable model. To find the variables which best predicts and explains the churn of a customer, variable selection was used. The final model presented high predictive ability and provided useful information on possible causes for the churn of customers.

  HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)