Development of forecasting models and software for online advertising optimization

Abstract

Denne Masteroppgaven beskriver og diskuterer utviklingen av en algoritme for optimalisering av annonsekostnader i digital markedsføring. API tilkoblinger, en fremskrivningsmodell for besøksverdi og en søkealgoritme ble kombinert for å lage en algoritme for optimalisering av annonsekostnader. Den valge tilstandsrommodellen for fremskrivning viser lovende resultater og algoritmen har blitt implementert som en fungerende prototype i Python.

This Master’s thesis describes and discusses the creation of an algorithm for optimization of ad spend in online digital advertising. API connections, a visitor value forecasting model and a basic search algorithm was combined to create an algorithm for optimization of ad spending. The chosen state space based forecasting model shows promising results and the algorithm has been implemented as a working prototype in Python.