Professor Erik Brynjolfsson and Shachar Reichman
Sponsor: Suruga Bank
We intend to develop a model that incorporates search query volumes into financial products (e.g. loans and deposits) demand predictions. We will train the model on public financial data (interest rates, income distribution, geographic location and etc). We will quantify the improvements offered by our model in predictions of overall sales and of the demand for certain financial product. Specifically, we will use three different techniques to automatically identify correlates among searched keywords: algorithmic methods, thesaurus, and crowdsourcing. Our plan is to compare the performance of these methods in general and for specific categories of search queries and create a hybrid model that will efficiently combine the best method for each type of term.