A comparative study of regression analysis with and without search query data as arepresentation of public opinion
Abstract: Stock prediction models using search query data is a modern phenomena and arelatively unexplored subject which potentially yields improvements to currentlyestablished prediction algorithms. This thesis will strive to improve an autoregressiveprediction model by analyzing concurrent search query data to conclude whether ornot taking such data into account will improve the prediction model. Multiplealternatives for sources of search query data has been analyzed and Google Trendswas concluded as the most suitable candidate. The thesis found no strong indicatorthat amplifying the autoregressive algorithm with Google Trends data would producebetter stock predictions. There remains to be found an elegant solution to improvingprediction models using Google Trends data.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)