Predictive capability of Google search in forecasting tourist arrival in the Philippines
Date
2017-12-22
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Google Trends, a new source of data, shows the index of worldwide search volume queries in the
internet. This study took advantage of this open data to look into its potentiality in improving the
tourist arrival forecast for the Philippines using fifty keywords, which were divided into three
categories: tourism, political and economic, and safety and order. A 12-month tourism forecast
horizon were made for the top five largest sources of tourists for the Philippines, including the
OFWs, using three models. A multivariate Vector Autoregression (VAR) model with Google data
and two benchmark models: Autoregressive Integrated Moving Average (ARIMA) and a VAR
model without Google data were compared using RMSE and MAPE criteria. This study provided
evidence that the VAR model with Google data generally performed better in terms of RMSE and
MAPE criteria as compared to the benchmark models. The set of keywords under the tourism
category also proved to be the most significant in predicting tourist arrivals.
Description
Keywords
Google trends, explanatory variables, Vector Autoregression Model (VAR), Autoregressive Integrated Moving Average (ARIMA), Root Mean Square Error (RMSE), Moving Average Percentage Error (MAPE)