Data Balancing to Improve Prediction of Project Success in the Telecom Sector

• Nuño Basurto ^[1] ; Alfredo Jiménez ^[2] ; Secil Bayraktar ^[3] ; Álvaro Herrero ^[1]
  1. [1] Universidad de Burgos

     Universidad de Burgos

     Burgos, España

     
  2. [2] Kedge Business School

     Kedge Business School

     Arrondissement de Bordeaux, Francia

     
  3. [3] Toulouse Business School

     Toulouse Business School

     Arrondissement de Toulouse, Francia

     

  Mostrar afiliaciones +
• Localización: 15th International Conference on Soft Computing Models in Industrial and Environmental Applications (SOCO 2020): Burgos, Spain ; September 2020 / coord. por Álvaro Herrero Cosío, Carlos Cambra Baseca, Daniel Urda Muñoz, Javier Sedano Franco, Héctor Quintián Pardo, Emilio Santiago Corchado Rodríguez, 2021, ISBN 978-3-030-57802-2, págs. 366-373
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • Investments in the telecom industry are often conducted through private participation projects, allowing a group of investors to build and/or opérate large infrastructure projects in the host country. As governments progressively removed the barriers to foreign ownership in this sector, these investment consortia have become increasingly international. Obviously, an accurate and early prediction of the success of such projects is very useful. Softcomputing can certainly contribute to address such challenge. However, the error rate obtained by classifiers when trying to forecast the project success is high due to the class imbalance (success vs. fail). To overcome such problem, present paper proposes the application of classifiers (Support Vector Machines and Random Forest) to data improved by means of data balancing techniques (both oversampling and undersampling). Results have been obtained on a real-life and publicly-available dataset from the World Bank.